Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Physiological roles of endogenous neurosteroids
at α2 subunit-containing GABAA receptors
Elizabeth Jane Durkin
A thesis submitted to University College London for the degree of Doctor
of Philosophy
October 2012
Department of Neuroscience, Physiology and Pharmacology
University College London
Gower Street
London
WC1E 6BT
Declaration 2
Declaration
I, Elizabeth Durkin, confirm that the work presented in this thesis is my own.
Where information has been derived from other sources, I confirm that this has
been indicated in the thesis
Abstract 3
Abstract
Neurosteroids are important endogenous modulators of the major inhibitory
neurotransmitter receptor in the brain, the γ-amino-butyric acid type A (GABAA)
receptor. They are involved in numerous physiological processes, and are linked to
several central nervous system disorders, including depression and anxiety. The
neurosteroids allopregnanolone and allo-tetrahydro-deoxy-corticosterone (THDOC)
have many effects in animal models (anxiolysis, analgesia, sedation, anticonvulsion,
antidepressive), suggesting they could be useful therapeutic agents, for example in
anxiety, stress and mood disorders.
Neurosteroids potentiate GABA-activated currents by binding to a conserved site within
α subunits. Potentiation can be eliminated by hydrophobic substitution of the α1Q241
residue (or equivalent in other α isoforms). Previous studies suggest that α2 subunits
are key components in neural circuits affecting anxiety and depression, and that
neurosteroids are endogenous anxiolytics. It is therefore possible that this anxiolysis
occurs via potentiation at α2 subunit-containing receptors. To examine this hypothesis,
α2Q241M knock-in mice were generated, and used to define the roles of α2 subunits in
mediating effects of endogenous and injected neurosteroids.
Biochemical and imaging analyses indicated that relative expression levels and
localization of GABAA receptor α1-α5 subunits were unaffected, suggesting the knock-
in had not caused any compensatory effects. Electrophysiological characterization of
cells in hippocampal and nucleus accumbens brain slices revealed faster-decaying
inhibitory synaptic transmission in α2Q241M mice. Furthermore, the response to applied
THDOC was markedly reduced compared to wild-type cells. α2 subunits therefore
formed a major component of synaptic GABAA receptors in these areas.
The α2Q241M knock-ins showed greater anxiety levels in two classical rodent anxiety
paradigms (light-dark box and elevated plus maze), consistent with endogenous
neurosteroids mediating anxiolysis via α2-type GABAA receptors. In addition, the
anxiolytic response to injected THDOC is impaired by the α2Q241M mutation, which
would identify the α2 isoform as an appropriate target for generating receptor subtype-
selective neurosteroid therapeutics for anxiety disorders.
Acknowledgements 4
Acknowledgements
I will forever be indebted to the many people who have helped me along my PhD
journey, especially to Prof. Trevor Smart, for allowing me to work on such a challenging
and exciting project, and for his unending support, guidance and enduring positivity.
This project would not have been possible without the financial support of the MRC,
and helpful advice from my thesis committee, Stuart Cull-Candy, Stephen Nurrish and
Antonella Riccio.
Thank you to Mike Lumb, for generating the transgenic mouse strain, together with
Steve Moss and GenOway. For generously providing lab space and loaning me the
equipment necessary for the behavioural analyses, I am grateful to Clare Stanford and
Stephen Hunt. I would also like to thank Clare, and members of her lab, especially
Ruth Weir and Ewelina Grabowska, for teaching me behavioural techniques, and
helping me interpret results from these experiments. Thank you also to Martin Stocker
for allowing access to his microtome.
Thanks to all members of the Smart lab, past and present, for creating an excellent
environment to work in, and being a great source of fun and of encouragement through
the inevitable experimental difficulties. I would particularly like to thank Damian and
Phil, for their excellent teaching and support, and for being on hand whenever my rig
misbehaved! Also, thanks to Saad for sharing his imaging expertise, and PhD pep
talks.
I would also like to take this opportunity to thank all of the Durkin ‘clan’ – above all,
Mum, Dad, Rob, Billy and Pat – for reminding me that there is life outside of my PhD,
and whose love and support has given me the drive to make it through my academic
career to date. I am particularly grateful to my PhD buddies who have shared in this
rollercoaster ride, especially Sinead and Kasia, who have always been there for a chat
over coffee.
Finally, the biggest of thanks is owed to Stuart for his patience, constant
encouragement, and excellent curries!
Contents 5
Contents
Declaration .................................................................................................................................................. 2
Abstract ....................................................................................................................................................... 3
Acknowledgements .................................................................................................................................... 4
List of Figures ............................................................................................................................................. 8
List of Tables............................................................................................................................................... 9
List of Abbreviations ................................................................................................................................ 10
Chapter 1: Introduction ............................................................................................................................ 12
1.1. GABAA receptors ............................................................................................................................... 12
1.1.1. GABAA receptor composition and function ............................................................................................ 13
1.1.2. GABAA receptor modulation: trafficking ................................................................................................. 18
1.1.3. GABAA receptor modulation: ligands and post-translational modification .............................................. 21
1.2. Neurosteroid physiology and pharmacology .................................................................................. 27
1.2.1. Endogenous neurosteroids: synthesis and roles ................................................................................... 27
1.2.2. Neurosteroid binding to GABAA receptors ............................................................................................. 30
1.2.3. Physiological modulation of GABAA receptors by neurosteroids ........................................................... 32
1.2.4. Neurosteroids: therapeutic potential ...................................................................................................... 34
1.3. Neurosteroids and GABAA receptors in anxiety ............................................................................. 35
1.3.1. The HPA axis and neurosteroids as endogenous anxiolytics ................................................................ 35
1.3.2. Defining the GABAA receptor α subunits that are important in anxiety and anxiolysis ........................... 37
1.4. Neurosteroids and GABAA receptors in depression ...................................................................... 39
1.4.1. Roles for GABAA receptors in depression ............................................................................................. 39
1.4.2. Roles for neurosteroids as antidepressants .......................................................................................... 43
1.4.3. The nucleus accumbens in reward and depression .............................................................................. 44
1.5. Thesis Aims........................................................................................................................................ 46
1.5.1. Generation of an α2Q241M knock-in mouse ............................................................................................. 46
1.5.2. Electrophysiological characterisation of α2Q241M mice ........................................................................... 46
1.5.3. Behavioural characterisation of α2Q241M mice ........................................................................................ 47
1.5.4. Summary of thesis aims ........................................................................................................................ 47
Chapter 2: Materials and Methods .......................................................................................................... 49
2.1. Materials ............................................................................................................................................. 49
2.1.1. Reagents .............................................................................................................................................. 49
2.1.2. Antibodies ............................................................................................................................................. 49
2.2. Animals ............................................................................................................................................... 50
2.2.1. Generating GABAA receptor α2Q241M mutant mice ................................................................................. 52
2.2.2. Breeding ............................................................................................................................................... 54
2.2.3. Genotyping ........................................................................................................................................... 54
2.3. Western blotting ................................................................................................................................ 55
2.3.1. Protein isolation .................................................................................................................................... 55
2.3.2. Polyacrylamide gel electrophoresis (PAGE) and blotting ...................................................................... 56
2.4. Immunofluorescence ......................................................................................................................... 58
2.4.1 Brain sectioning for immunofluorescence .............................................................................................. 58
2.4.2. Staining sectioned tissue ...................................................................................................................... 59
2.4.3. Image acquisition and analysis ............................................................................................................. 60
Contents 6
2.5. HEK293 cell culture and electrophysiology .................................................................................... 63
2.5.1. HEK293 cell culture .............................................................................................................................. 63
2.5.2. HEK293 cell transfection ....................................................................................................................... 63
2.5.3. HEK293 cell electrophysiology .............................................................................................................. 64
2.6. Brain slice electrophysiology ........................................................................................................... 66
2.6.1. Preparation of slices ............................................................................................................................. 66
2.6.2. Whole-cell patch clamp recording ......................................................................................................... 67
2.6.3. Analysis of spontaneous inhibitory post-synaptic currents (sIPSCs) ..................................................... 68
2.6.4. Analysis of tonic GABA currents ........................................................................................................... 69
2.7. Behavioural Analyses ....................................................................................................................... 70
2.7.1. Animal handling and drug administration .............................................................................................. 70
2.7.2. Elevated plus maze ............................................................................................................................... 70
2.7.3. Light-Dark Box ...................................................................................................................................... 72
2.7.4. Tail-suspension test .............................................................................................................................. 73
2.7.5. Rotarod ................................................................................................................................................. 75
2.8. Statistics ............................................................................................................................................. 77
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA
receptor α2 subunit .................................................................................................................................. 80
3.1. Introduction ........................................................................................................................................ 80
3.2. Results ................................................................................................................................................ 83
3.2.1. Confirming that α2Q241M has no effect on GABA and diazepam sensitivity of αβγ receptors .................. 83
3.2.2. Probing the effect of α2Q241M on responses to both THDOC and diazepam .......................................... 86
3.2.3. Generating an α2Q241M knock-in transgenic mouse ............................................................................... 88
3.2.4. α2Q241M has no effect on the expression of GABAA receptor subunits α1-α4 in Western blot assays ..... 88
3.2.5. α2Q241M has no effect on the expression of GABAA receptor subunits α1-α5 in quantitative
immunofluorescence assays ........................................................................................................................... 90
3.3. Discussion........................................................................................................................................ 100
3.3.1. α2Q241M and GABAA receptor function ................................................................................................. 100
3.3.2. No compensatory changes in GABAA receptor α subunit expression in α2Q241M mice ......................... 101
3.3.3. Endogenous neurosteroids may modulate the in vivo response to other GABAA receptor potentiators
..................................................................................................................................................................... 105
3.4. Conclusions ..................................................................................................................................... 106
Chapter 4: Electrophysiological consequences of losing neurosteroid modulation at GABAA
receptor α2 subunits .............................................................................................................................. 107
4.1. Introduction ...................................................................................................................................... 107
4.1.1. GABAergic neurotransmission in the hippocampus ............................................................................ 107
4.1.2. GABAergic neurotransmission in the nucleus accumbens .................................................................. 110
4.1.3. Modulation of synaptic and tonic GABAergic currents by benzodiazepines and neurosteroids ........... 111
4.2. Results .............................................................................................................................................. 113
4.2.1. α2Q241M alters baseline inhibitory neurotransmission in acute slices of hippocampus and nucleus
accumbens ................................................................................................................................................... 113
4.2.2. α2Q241M specifically diminishes neurosteroid potentiation of synaptic inhibitory neurotransmission ..... 116
4.2.3. α2Q241M specifically diminishes neurosteroid potentiation of tonic inhibitory neurotransmission in dentate
gyrus granule cells ........................................................................................................................................ 124
4.3. Discussion........................................................................................................................................ 127
Contents 7
4.3.1. Effects of α2Q241M on baseline synaptic and tonic GABA transmission ................................................ 127
4.3.2. α2Q241M specifically diminishes neurosteroid potentiation of synaptic transmission ............................. 129
4.3.3. α2Q241M reveals a role for α2-type GABAA receptors in tonic currents .................................................. 132
4.3.4. Limitations of the electrophysiological characterisation ....................................................................... 132
4.4. Conclusions ..................................................................................................................................... 134
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M
knock-ins ............................. 136
5.1. Introduction ...................................................................................................................................... 136
5.1.1. Behaviour of α2Q241M mice defines the physiological and pharmacological roles of neurosteroids acting
at GABAA receptor α2 subunits ..................................................................................................................... 136
5.1.2. GABAA receptor α2 subunits and neurosteroids in anxiety .................................................................. 137
5.1.3. GABAA receptor α2 subunits and neurosteroids in depression ............................................................ 138
5.2. Results .............................................................................................................................................. 140
5.2.1. Endogenous neurosteroids act via GABAA receptor α2 subunits to determine basal anxiety levels .... 140
5.2.2. GABAA receptor α2 subunits mediate the anxiolysis of injected neurosteroids at low doses ............... 143
5.2.3. Anxiolytic effects of pentobarbital and diazepam are retained in α2Q241M knock-in mice ...................... 149
5.2.4. Knock-in mutation α2Q241M does not confer a depression phenotype .................................................. 155
5.2.5. Ataxic effects of THDOC injection in hom mice: not a consequence of increased motor impairment .. 156
5.3. Discussion........................................................................................................................................ 158
5.3.1. Neurosteroids act via α2-type GABAA receptors to modulate anxiety state ......................................... 158
5.3.2. Anxiolytic effects of pentobarbital and diazepam ................................................................................ 162
5.3.3. Depression may not involve neurosteroids acting at α2-type GABAA receptors .................................. 164
5.3.4. Exploring the mechanism underlying THDOC’s ataxic effects in α2Q241M mice .................................... 166
5.4. Conclusions ..................................................................................................................................... 169
Chapter 6: General Discussion.............................................................................................................. 171
6.1. Endogenous neurosteroids: functions at α2-type GABAA receptors.......................................... 171
6.1.1. Anxiolysis but not anti-depression in unperturbed state ...................................................................... 171
6.1.2. α2Q241M knock-in has no effect on receptor expression........................................................................ 172
6.1.3. Contribution of α2-type GABAA receptors to phasic and tonic transmission ........................................ 172
6.2. The therapeutic potential of neurosteroids in anxiety and depression ...................................... 173
6.2.1. THDOC is not antidepressant ............................................................................................................. 173
6.2.2. α2-type GABAA receptors are an appropriate target for neurosteroid anxiolytics ................................ 175
6.3. Remaining questions and future work ........................................................................................... 175
6.3.1. Screening for compensatory changes in α2Q241M mice ........................................................................ 175
6.3.2. Understanding the activity effects ....................................................................................................... 177
6.3.3. Future work with α2Q241M mutant mice ................................................................................................. 177
6.4. Concluding statement ..................................................................................................................... 179
References .............................................................................................................................................. 180
List of Figures 8
List of Figures
Figure 1.1 – GABAA receptor subunit assembly and modulator binding sites ............................................. 14
Figure 1.2 – GABAA receptor binding partners .......................................................................................... 19
Figure 1.3 – Neurosteroid synthesis pathways .......................................................................................... 28
Figure 1.4 – GABAA receptor subunit assembly and neurosteroid binding sites ........................................ 31
Figure 2.1 – Generation of the mutant mice .............................................................................................. 53
Figure 2.2 – Optimising conditions for Western Blotting ............................................................................ 58
Figure 2.3 – FociPicker3D successfully identifies immunopositive punctae if images lack diffuse staining 62
Figure 2.4 – Schematic representation of the Y-tube ................................................................................. 65
Figure 2.5 – The elevated plus maze ......................................................................................................... 71
Figure 2.6 – The light-dark box .................................................................................................................. 73
Figure 2.7 – Tail suspension test equipment ............................................................................................. 75
Figure 2.8 – Rotarod protocol .................................................................................................................... 77
Figure 3.1 – Mutation α2Q241M
has no effect on the GABA or diazepam sensitivity of α2β3γ2S receptors
expressed in HEK293 cells ..................................................................................................... 85
Figure 3.2 – Mutation α2Q241M
specifically abolishes the potentiating effect of THDOC on GABA EC15
responses of α2β3γ2S receptors expressed in HEK293 cells ................................................. 87
Figure 3.3 – No change in expression of GABAA subunits α1-α4 .............................................................. 89
Figure 3.4 – Immunofluorescent localisation of GABAA subunits α1-α5 in coronal sections of hippocampus
and nucleus accumbens ........................................................................................................ 92
Figure 3.5 – Quantitation of immunofluorescent staining for GABAA subunits α1-α5 in coronal sections of
hippocampus and nucleus accumbens .................................................................................. 97
Figure 3.6 – Quantitation of immunofluorescent punctae for GABAA subunits α3 in the dentate gyrus and
α1 in the nucleus accumbens ................................................................................................. 99
Figure 4.1 – Hippocampal networks relevant to anxiety and depression ................................................. 109
Figure 4.2 – α2Q241M
speeds the decay of baseline IPSCs ...................................................................... 114
Figure 4.3 – THDOC effects on sIPSC decay time in CA1 pyramidal cells are diminished in recordings from
homozygous knock-ins ........................................................................................................ 118
Figure 4.4 – Homozygous α2Q241M
specifically disrupts neurosteroid potentiation of sIPSCs in DG GCs and
NAcc MSNs .......................................................................................................................... 121
Figure 4.5 – Homozygous mutation α2Q241M
specifically disrupts neurosteroid potentiation of tonic currents
in DG GCs ............................................................................................................................ 125
Figure 5.1 – α2Q241M
confers an anxious phenotype in the elevated plus maze ...................................... 141
Figure 5.2 – α2Q241M
confers an anxious phenotype in the light-dark box ................................................ 142
Figure 5.3 – α2Q241M
influences behavioural responses to injected THDOC on the elevated plus maze . 144
Figure 5.4 – α2Q241M
alters response to THDOC injection in the light-dark box ....................................... 146
Figure 5.5 – Time-course analysis of light-dark box results ..................................................................... 148
Figure 5.6 – α2Q241M
does not influence pentobarbital’s effect on elevated plus maze behaviour ........... 151
Figure 5.7 – α2Q241M
does not influence diazepam’s effect on elevated plus maze behaviour ................ 152
Figure 5.8 – α2Q241M
alters response to diazepam in the light-dark box ................................................... 154
Figure 5.9 – Immobility in the tail suspension test is unaffected by mutation α2Q241M
.............................. 155
Figure 5.10 – α2Q241M
does not affect performance on the accelerating rotarod under baseline or THDOC-
treated conditions ................................................................................................................. 157
List of Tables 9
List of Tables
Table 1.1 – Subunit-specific actions of benzodiazepines ........................................................................... 22
Table 1.2 – Phosphorylation sites on GABAA receptors ............................................................................ 24
Table 2.1 – Details of antisera employed in this project ............................................................................. 51
Table 2.2 – Amount of total protein loaded for Western blotting ................................................................ 57
Table 3.1 – Hill fit parameters for GABA and diazepam responses ........................................................... 84
Table 4.1 – Comparing baseline sIPSCs recorded from acute brain slices ............................................. 115
Table 4.2 – Comparing baseline tonic currents in slice recordings .......................................................... 116
Table 4.3 – THDOC-induced changes CA1 PC sIPSC parameters ......................................................... 119
Table 4.4 – Changes in DG GC sIPSC parameters ................................................................................. 122
Table 4.5 – Changes in NAcc MSN sIPSC parameters ........................................................................... 123
Table 4.6 – Comparing changes in tonic currents after equilibration with THDOC or diazepam ............. 126
List of Abbreviations 10
List of Abbreviations
3α-HSD, 3β-HSD 3α-hydroxy-steroid dehydrogenase, 3β-hydroxy-steroid dehydrogenase
aCSF Artificial cerebrospinal fluid
AMPA 2-amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl)propanoic acid
ANOVA Analysis of variance
AP2 Clathrin adaptor protein 2
ASPA Animals (Scientific Procedures) Act, 1986
BCA Bicinchoninic acid
BDNF Brain-derived neurotrophic factor
CA1 – CA3 Cornu ammonis (CA) regions 1-3 of the hippocampus
CaMKII Ca2+
/calmodulin-dependent kinase II
cDNA Complementary deoxy-ribonucleic acid
CNS Central nervous system
CSF Cerebrospinal fluid
d.f. Degrees of freedom
DG Dentate gyrus
DHEA Dehydro-epi-androsterone
DMEM Dulbecco’s modified Eagle medium
DMSO Dimethyl sulphoxide
EC15 The concentration of substance producing a response 15% of the maximal
EC50 The concentration of substance producing a response 50% of the maximal
EDTA Ethylene-diamine-tetra-acetic acid
EEG Electro-encephalogram
eGFP Enhanced green fluorescent protein
ES cell Embryonic stem cell
GABA Gamma amino butyric acid
GABAA, GABAB, GABAC GABA receptor subclasses A-C
GABARAP GABAA-receptor-associated protein
GAD65, GAD67 Glutamate decarboxylases 65 and 67
GC Granule cell
HAP-1 Huntingtin-associated protein 1
HBSS Hank’s balanced salt solution
HEK293 Human embryonic kidney 293
HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
Het Heterozygous, heterozygote
Hom Homozygous, homozygote
HPA axis Hypothalamic-pituitary-adrenal axis
HPLC High-performance liquid chromatography
HRP Horseradish peroxidase
I.E.I. Inter-event interval
I.F. Immunofluorescence
ICSS Intra-cranial self-stimulation
IPSC Inhibitory postsynaptic current
kb Kilo-base-pairs
kDa Kilo-Daltons
KIF5 Kinesin family motor protein 5
LSD Least significant difference
List of Abbreviations 11
Lux Light intensity measure, equivalent to lumens per square metre
M1 – M4 Transmembrane helices 1 - 4
MF Mass fragmentography
mIPSC Miniature inhibitory postsynaptic current
mRNA Messenger ribonucleic acid
MSN Medium spiny neuron
NA Numerical aperture
NAcc Nucleus Accumbens
NMDA N-Methyl-D-aspartate
P450c17 17α hydroxylase, c17,20 lyase
P450scc Cytochrome P450 cholesterol side-chain cleavage
PAGE Poly-acrylamide gel electrophoresis
PB Phosphate buffer
PBS Phosphate buffered saline
PC Pyramidal cell
PCR Polymerase Chain Reaction
PKA, PKB, PKC, PKG Protein kinases A, B, C and G
PLIC1 Protein that links integrin-associated protein with the cytoskeleton-1
PP1, PP2A Protein phosphatases 1 and 2A
PRIP Phospholipase-C related inactive protein
Pxx (e.g. P18) Postnatal age xx days
r.m.s. Root mean square
r.p.m. Rotations per minute
RACK1 Receptor for activated C kinase1
Raft1 Rapamycin and FKB12 target protein
REM sleep Rapid eye movement sleep
ROI Region of interest
Rs Series resistance
RT Room temperature
s.e.m. Standard error of the mean
SDS Sodium-dodecyl sulphate
sIPSC Spontaneous inhibitory postsynaptic current
SON Supraoptic nucleus
SPECT Single photon emission computed tomography
SSRI Selective serotonin reuptake inhibitor
StAR Steroidogenic acute regulatory protein
THDOC Allo-tetrahydro-deoxy-corticosterone (3α,21-dihydroxy-5α-pregnan-20-one)
TSPO 18 kDa translocator protein, also called ‘peripheral benzodiazepine receptor’
Veh Vehicle
VTA Ventral tegmental area
WB Western blot
Wt Wild-type
γ2L, γ2S Long (L) and short (S) isoforms of the γ2 subunit
Chapter 1: Introduction 12
Chapter 1: Introduction
Neurosteroid modulation of GABAergic neurotransmission in physiology
and pathophysiology
1.1. GABAA receptors
Gamma-aminobutyric acid (GABA) is the major inhibitory neurotransmitter in the
mammalian central nervous system (CNS), playing a central role in regulating
neuronal excitability. GABA fulfils this role by activating two classes of receptor:
ionotropic type-A receptors (GABAA receptors) and metabotropic type-B
(GABAB) receptors. The focus of this study is the GABAA receptor family, which
encompasses Cl- and bicarbonate-permeable channels of the Cys-loop-
containing ligand-gated ion channel family (of which the nicotinic acetylcholine
receptor is the founding member) (Barnard et al., 1998), now known as the
pentameric ligand-gated ion channels (Miller & Smart, 2010; Corringer et al.,
2012).
Excitatory and inhibitory neurotransmission are finely balanced processes, and
any disruption of this balance can be detrimental to brain function. Improper
GABAergic signalling is seen in neurodegenerative diseases such as
Huntington’s, after ischemic episodes, and also in epilepsy, anxiety disorders,
depression and schizophrenia (Fritschy & Brunig, 2003; Mohler, 2006b). It is
therefore unsurprising that GABAA receptors represent a major pharmacological
target for the treatment of these disorders. This thesis will focus on the roles of
GABAA receptors in anxiety and depression, with particular attention paid to
neurosteroids as players in the disease process, as well as assessing their
potential as therapeutic agents for these disorders.
Chapter 1: Introduction 13
1.1.1. GABAA receptor composition and function
The GABAA-receptor gene family encompasses several subunits, some of
which are present in multiple isoforms (α1-6, β1-3, γ1-3, δ, ε, π and ρ1-3
(Barnard et al., 1998; Korpi et al., 2002)), with further diversity imparted by
alternative splicing (e.g. γ2 is expressed as short and long splice forms – γ2S
and γ2L, respectively (Whiting et al., 1990; Kofuji et al., 1991; Glencorse et al.,
1992)). When studied in recombinant expression systems, functional receptors
may assemble from a single subunit (e.g. β or ρ), or two subunits (e.g. α with β).
However, the majority of native receptors in vivo are thought to be pentamers of
2α, 2β and 1x subunit (Fig. 1.1), where x is typically a γ subunit in synaptic
receptors, but could be δ in receptors outside the synapse (Brickley et al., 1999;
Mody, 2001; Moss & Smart, 2001; Sieghart & Sperk, 2002; Mohler, 2006a).
These rules are not strict, however, since typically synaptic subunits are found
outside of synapses too (e.g. α1 and γ2 in electron micrographs of cerebellar
granule cells (Nusser et al., 1998)), and αβ receptors (i.e. without γ or δ
subunits) have been observed in hippocampal neurons (Mortensen & Smart,
2006). The γ subunit in αβγ receptors may be replaced by ε or π, whilst θ
subunits are thought to take the place of β subunits (Sieghart & Sperk, 2002).
Expression of subunits ρ1-3 is mostly restricted to the retina, where they exist
as homo- or hetero-oligomers with properties distinct from αβx assemblies,
leading to their separate classification as GABAC receptors (Sieghart & Sperk,
2002).
Chapter 1: Introduction 14
Figure 1.1 – GABAA receptor subunit assembly and modulator binding sites
Each GABAA receptor subunit has the depicted topology (A) of a large extracellular
amino-terminal (NH2) domain, four transmembrane helices (M1-4), and a short
extracellular carboxy-terminus (COOH) (Korpi et al., 2002). The large intracellular loop
between M3 and M4 helices is a key site for protein-protein interactions and
phosphorylation, both of which are means for regulating receptor location and function
(see main text). Subunits are thought to assemble as pentamers of stoichiometry
2α:2β:1x subunits (Chang et al., 1996; Tretter et al., 1997; Farrar et al., 1999). The
proposed arrangement of subunits is depicted as viewed from the extracellular surface
(B) or in the plane of the membrane (C). The locations of binding sites for GABA,
benzodiazepines (BDZ, note that this site depends on x being a γ subunit) and
neurosteroid activation (Nster (a)) and neurosteroid potentiation (Nster (p)) are
indicated (Korpi et al., 2002; Hosie et al., 2006; Hosie et al., 2009).
Given the large number of subunit isoforms, there are theoretically many
permutations of αβx assemblies. However, far fewer combinations are actually
found in vivo, depending on which subunits are co-expressed within a neuron
(McKernan & Whiting, 1996; Sieghart & Sperk, 2002; Mohler, 2006a). Distinct
expression profiles are observed for each subunit across brain areas, which
alter with development (Laurie et al., 1992a; Laurie et al., 1992b; Wisden et al.,
1992; Pirker et al., 2000). Furthermore, within a particular brain region and/or
cell type, there can be cell-to-cell variations in receptor expression (e.g. using
single-cell polymerase chain reaction (PCR) measurements of messenger
ribonucleic acid (mRNA) levels, some cerebellar granule cells appear to
express only α1, others only α6, and a third subset express both (Santi et al.,
1994)). The likely receptor subunit combinations formed in vivo have been
Chapter 1: Introduction 15
defined by examining protein co-distribution with immunostaining and mRNA
co-distribution by in situ hybridisation (Somogyi et al., 1989; Fritschy et al.,
1992; Wisden & Seeburg, 1992), and co-immunoprecipitation of subunits from
brain tissue (Benke et al., 1991; McKernan & Whiting, 1996). The two most
abundant receptor combinations in rat brain are thought to be α1β2γ2 (40-60%)
and α2β2/3γ2 (15-20%); other less common combinations include α3βnγ2 (10-
15%), α4βnγ2/α4βnδ (<5%), α5βnγ2 (<5%) and α6βnγ2/α6βnδ (<5%)
(McKernan & Whiting, 1996; Sieghart & Sperk, 2002; Mohler, 2006a). It is also
possible that the two copies of α or β subunits within a receptor can be different
isoforms – e.g. α1α3β2/3γ2 receptors (Fritschy et al., 1992).
The agonist-gated pore of GABAA receptors allows transmembrane passage of
Cl- and HCO3- ions; the direction of the ion flow is determined by their
transmembrane electrochemical gradients, which can be developmentally
regulated (switching from excitatory to inhibitory during brain development (Ben-
Ari, 2002)). Depending on their properties and subcellular location, GABAA
receptors can carry two types of current: ‘phasic’ and ‘tonic’ (Mody, 2001;
Semyanov et al., 2004; Farrant & Nusser, 2005). Phasic events, or inhibitory
post synaptic currents (IPSCs), involve action-potential stimulated release of
GABA at the synaptic cleft, which stimulates opening of synaptically-located
GABAA receptors; events are short-lived, due to rapid GABA clearance from the
cleft. In contrast, tonic currents result from more persistent activation of GABAA
receptors, which are responding to the low ambient GABA levels outside the
synapse (Mody, 2001; Semyanov et al., 2004; Farrant & Nusser, 2005). Both
types of inhibition can be considered a means to counter excessive neuronal
network excitation; phasic currents are also involved in generating rhythmic
network activities, whilst tonic inhibition can modulate the input-output
relationship for excitatory signalling onto a particular cell (Semyanov et al.,
2004; Farrant & Nusser, 2005).
The source of ambient GABA responsible for tonic currents varies, but can
involve spillover of synaptically released GABA, reverse transport at synapses
and non-synaptic sources of GABA (including release from astrocytes)
Chapter 1: Introduction 16
(Richerson, 2004; Semyanov et al., 2004; Farrant & Nusser, 2005). The
importance of vesicular release is supported by the reduction in extracellular
GABA concentrations after blocking action potentials with tetrodotoxin (Bianchi
et al., 2003; Xi et al., 2003) and also by the strong positive correlation between
phasic and tonic current amplitudes recorded from brain slices under a range of
conditions (Glykys & Mody, 2007). Nevertheless, the residual measures of
extracellular GABA after tetrodotoxin treatment of the hippocampus (Bianchi et
al., 2003) or nucleus accumbens (Xi et al., 2003) indicate that as much as 75%
of extracellular GABA in these regions is of non-vesicular origin.
In order to pass current during prolonged exposure to low ambient GABA
concentrations, tonic-carrying receptors must have a higher GABA affinity and
slower desensitisation kinetics than those involved in phasic currents (Farrant &
Nusser, 2005). Approaches to determine the identity of tonic vs. phasic
receptors include imaging (to demonstrate appropriate subcellular localisation),
pharmacology (demonstrating current sensitivity to subunit-selective
compounds) and subunit knock-outs (loss of current when subunit x is ablated).
The results of these experiments are not always unequivocal, probably because
the identity of receptors passing each type of current varies across cell types,
and within a particular cell type according to the recording conditions employed.
This is perhaps best illustrated by considering the reports for tonic currents in
hippocampal pyramidal cells (PCs), particularly those in the cornu ammonis 1
(CA1) region. Semyanov et al. (2003) found no detectable tonic current in PCs
from guinea pig hippocampi at baseline, and required inhibition of GABA uptake
to reveal a tonic current. In contrast, Prenosil et al. (2006) observe a clear tonic
current in CA1 PCs without manipulating extracellular GABA levels. Where a
tonic current has been detected in wild-type CA1 PCs, pharmacological profiling
pointed to roles for α5, β2/3 and γ2, whilst suggesting no involvement of α4, α6,
ε or δ subunits (Caraiscos et al., 2004). In addition, knock-out of α5 subunits
diminished tonic current recorded from CA1 (Caraiscos et al., 2004). However,
a role for α5 subunits is not universally supported: for example, Prenosil et al.
(2006) find their tonic current to be insensitive to L655-708, an α5-selective
benzodiazepine-site inverse agonist. A role for δ subunits in tonic currents
seemed unlikely on the basis of several observations: δ subunit knock-out mice
Chapter 1: Introduction 17
(δ-/- mice) have undiminished tonic currents in hippocampal pyramidal cells in
culture (Caraiscos et al., 2004) and brain slice tissue (Stell et al., 2003; Glykys
et al., 2008), and the tonic current of wild-type cells is insensitive to the
neurosteroid THDOC (allo-tetrahydro-deoxy-corticosterone) (Stell et al., 2003).
However, they cannot rule out compensatory alterations in response to losing
the δ subunit, nor does THDOC insensitivity necessarily imply a lack of δ
subunit contribution. Indeed, a role for δ subunits has more recently been
acknowledged, because the residual tonic current in CA1 PCs of α5-/- mice is
lost in double knockout α5-/-δ-/- mice (Glykys et al., 2008). Interestingly,
treatment of slices from δ-/- mice with L-655,708 does not abolish the tonic
current (Glykys et al., 2008), suggesting yet more subunits could contribute to
tonic currents in CA1 PCs. Indeed, there is evidence for a contribution of αnβn
combinations to tonic currents of cultured hippocampal pyramidal cells
(Mortensen & Smart, 2006). Data therefore predict a dominant role for α5β2/3γ2
receptors in determining CA1 PC tonic current, but do not rule out contributions
from other subunit combinations. On the basis of results from similar
experimental approaches, cerebellar granule cell tonic current is believed to be
mostly mediated by α6βnδ receptors (Brickley et al., 2001), and that in dentate
gyrus granule cells (DG GC) by α4βnδ (Nusser & Mody, 2002; Stell et al.,
2003).
Regardless of the identity of receptors passing GABAergic currents, it is
appreciated that this neurotransmission is highly plastic, and is modulated by a
number of physiological and pharmacological processes (Luscher & Keller,
2004). Mechanisms exist to alter the density of receptors in a given membrane
region, as well as the amount of current that flows through these receptors in
response to GABA binding. Some of these mechanisms are described in the
following sections (1.1.2 and 1.1.3).
Chapter 1: Introduction 18
1.1.2. GABAA receptor modulation: trafficking
The regulation of synaptic GABAA receptor trafficking and the roles of its many
receptor-associated partners in these processes has been extensively reviewed
elsewhere (Kittler & Moss, 2001; Moss & Smart, 2001; Fritschy & Brunig, 2003;
Kittler & Moss, 2003; Luscher & Keller, 2004; Arancibia-Carcamo & Kittler,
2009; Luscher et al., 2011b). A full discussion of these mechanisms is beyond
the scope of this thesis, but some of the key components (shown in Fig. 1.2) will
be described below. The majority of the GABAA receptor binding partners
associate with the intracellular domain between transmembrane helices M3 and
M4, and so it is unlikely that all of these interactions will take place at once.
Complexes are likely to be dynamic and transient in nature, allowing regulation
of receptor actions at particular locations and under specific conditions.
Gephyrin and γ2 subunits appear to be interdependent for synaptic clustering of
GABAA receptors (Essrich et al., 1998; Fischer et al., 2000), although the
interaction with gephyrin occurs via α subunits (Tretter et al., 2008; Tretter et al.,
2011). There are also gephyrin-independent means of receptor clustering, such
as interactions with dystrophin (Knuesel et al., 1999). Gephyrin’s roles probably
extend beyond synaptic anchoring of GABAA and glycine receptors, because it
interacts with numerous binding partners that may regulate cytoskeleton
dynamics, local protein translation and receptor trafficking (Fig. 1.2).
Chapter 1: Introduction 19
Figure 1.2 –GABAA receptor binding partners
GABAA receptors are modulated by interactions with a host of other proteins, some of
which are discussed in the main text. Receptor delivery to the cell surface involves
trafficking proteins GABARAP (GABAA receptor associated protein), KIF5 (kinesin
family motor protein 5) and PRIP (phospholipase-C related inactive protein), whilst
endocytosis for recycling or degradation involves proteins AP2 (clathrin adaptor protein
2) and PLIC1 (protein that links integrin-associated protein with the cytoskeleton-1).
Synaptic clustering of receptors involves interaction with dystrophin or with gephyrin,
which itself interacts with a number of proteins. Through these interactions, gephyrin
may bind to and regulate the cytoskeleton, as well as modulate protein translation via
Raft1 (rapamycin and FKB12 target protein). GABAA receptors can also bind proteins
involved in modulating receptor phosphorylation state, such as RACK1 (receptor for
activated C kinase1), PKCβII (protein kinase C, isoform βII) and PRIP. Phosphorylation
serves to regulate GABAA receptor function and trafficking, and may do so by
modulating receptor interactions with its binding partners (see main text).
Receptor targeting mechanisms are more sophisticated than simply directing to
a synaptic or extrasynaptic site: specific receptors can be localised at particular
synapses within a neuron. For example, within hippocampal pyramidal cells, α1
subunits are uniformly distributed, but α2 subunits are concentrated at synapses
local to the axon initial segment (Nusser et al., 1996). Targeting may depend on
signals from the presynaptic cell, because distinct interneurons synapse onto
Chapter 1: Introduction 20
these different regions of the principal cell: α1 subunits are at synapses with
parvalbumin-positive basket cells, whilst α2 subunits are clustered at synapses
from cholecystokinin-positive basket cells and parvalbumin-positive chandelier
cells (Nyiri et al., 2001; Luscher & Keller, 2004; Mohler, 2006a)). Subcellular
targeting has been probed using an artificial synapse model, which
demonstrates that as-yet undefined determinants within α2 vs. α6 subunits
direct them to the appropriate location (Wu et al., 2012). Indeed, ectopically-
over-expressed α6 subunits can direct receptors to extrasynaptic sites within
neurons, and ‘dominates’ over γ2 subunits within the same receptor in this
targeting (Wisden et al., 2002).
Although anchored by gephyrin and/or dystrophin, synaptic GABAA receptors
are by no means static entities. Receptors are constantly trafficking into and out
of the synaptic zone by membraneous transport (exo- and endocytosis), as well
as by lateral mobility within the membrane. Receptor internalisation involves
interaction of β and/or γ subunits with the clathrin adaptor protein, AP2,
depending on their phosphorylation state (Kittler et al., 2000; Kittler et al., 2005;
Luscher et al., 2011a). After internalisation, receptors may either recycle back to
the cell surface, or are targeted to the lysosome for degradation. The latter
route is favoured by lysine ubiquitination of the γ subunit (Arancibia-Carcamo et
al., 2009), whilst cell surface re-delivery is favoured by interaction of PLIC1 with
α and β subunits (Bedford et al., 2001). Cell surface delivery of GABAA
receptors is also facilitated by interaction of β subunits with PRIP (Kanematsu et
al., 2006; Mizokami et al., 2007) and the kinesin family motor protein, KIF5 (via
adaptor protein Huntingtin-associated protein 1, HAP1 (Twelvetrees et al.,
2010)). Trafficking from the Golgi apparatus to the cell surface also involves
interaction of γ subunits with GABARAP and with tubulin (the latter interaction
may occur via GABARAP and/or through HAP1-KIF5 complexes) (Item &
Sieghart, 1994; Wang et al., 1999; Kneussel et al., 2000; Kittler et al., 2001).
The cycling of GABAA receptors into and out of the cell surface membrane
appears to be occurring constitutively, because disruption of either process can
influence both the cell surface expression of γ-subunit containing GABAA
receptors, and synaptic GABAergic currents. For example, over a short
timescale (minutes) disruption of clathrin-mediated endocytosis increases
Chapter 1: Introduction 21
synaptic current amplitudes (Kittler et al., 2000); conversely, impairing cell
surface delivery by disrupting the KIF5-HAP1 interaction reduces current
amplitudes (Twelvetrees et al., 2010). Furthermore, despite their interaction with
anchoring proteins, synaptic GABAA receptors also show lateral mobility within
the membrane (Thomas et al., 2005). Movement of receptors into and out of
synaptic sites by this mechanism may represent the major means of altering
synaptic strength, and may occur over faster timescales than vesicular transport
mechanisms (Thomas et al., 2005).
1.1.3. GABAA receptor modulation: ligands and post-translational modification
A number of endogenous and exogenous ligands are allosteric modulators or
direct agonists of GABAA receptors including barbiturates, benzodiazepines,
neurosteroids and Zn2+ (details below). GABAA receptor activity can also be
modulated by direct modifications of the channel, including phosphorylation
(details below), protonation (Huang & Dillon, 1999; Wilkins et al., 2002, 2005)
and redox modifications of cysteine residues (Amato et al., 1999; Pan et al.,
2000).
Pharmacological modulation by barbiturates and benzodiazepines
The effects of barbiturates on GABAA receptor currents depends on their
concentration: potentiation, direct activation and inhibition occur with increasing
concentration (e.g. pentobarbitone shows all three responses, with efficacies
dependent on receptor subunit composition (Thompson et al., 1996)). By
increasing GABAA receptor currents, these compounds are effective as
anxiolytics and hypnotics. However, clinical barbiturate use has now been
superseded by that of benzodiazepines, which are also allosteric potentiators of
GABAA receptors, but are safer in overdose (Smith & Rudolph, 2012).
Responses to classical benzodiazepines (e.g. diazepam) require the presence
of a γ subunit and absence of the benzodiazepine-insensitive α4 or α6 isoforms
Chapter 1: Introduction 22
(Pritchett et al., 1989; Wieland et al., 1992). Their numerous behavioural effects
have been linked to their action at distinct GABAA receptor α subunit isoforms
(Table 1.1). Targeting benzodiazepines to specific α subunits may therefore
produce specific behavioural effects. Particular efforts have focussed on
generating α2-targeting benzodiazepines as non-sedating anxiolytics (see
Section 1.3.2).
Isoform Benzodiazepine effect References
α1 sedation, amnesia, anticonvulsion (Rudolph et al., 1999; McKernan et
al., 2000)
α2 anxiolysis and myorelaxation (Low et al., 2000; Crestani et al.,
2001)
α3 anxiolysis1 and myorelaxation
(Crestani et al., 2001; Dias et al.,
2005)
α5 sedative tolerance (van Rijnsoever et al., 2004)
Table 1.1 – Subunit-specific actions of benzodiazepines
There are six α subunit isoforms, and whilst all are neurosteroid sensitive (Hosie et al.,
2009), α4 and α6 subunits preclude benzodiazepine response due to a single amino
acid substitution (H to R) in the otherwise conserved binding site (Wieland et al., 1992).
Introducing this substitution into the benzodiazepine-responsive subunits (i.e. α1H101R,
α2H101R, α3H126R or α5H105R) abolishes their response to benzodiazepines in recombinant
systems, but leaves other properties of the channel unaffected (Wieland et al., 1992;
Benson et al., 1998). This table summarises the results of characterising knock-in mice
with each of these point mutations, which allowed dissection of the isoforms
responsible for the various behavioural effects of benzodiazepines.
1Note that α3H126R mice did not lack anxiolytic response to diazepam (Low et al., 2000),
but a role for this subunit was supported by observations of anxiolysis with an α3-
selective benzodiazepine (Dias et al., 2005)
Endogenous modulation by Zn2+
Experimentally, transmission through GABAA receptors can be modulated by
Zn2+ ions either directly (inhibitory effects on subunit combinations lacking γ2
subunits) or indirectly (by enhancing synaptic release of GABA in the
Chapter 1: Introduction 23
hippocampus) (Hosie et al., 2003; Smart et al., 2004). Histological stains for
Zn2+ demonstrate a strong presence in the telencephalon, hippocampus,
cerebral cortex and amygdala, suggesting that Zn2+ modulation occurs
endogenously (Smart et al., 2004). Chelation of Zn2+ ions increased the
amplitude of CA3 pyramidal cell IPSCs evoked by stimulation of mossy fiber
inputs, confirming a role for baseline endogenous Zn2+ in modulating
GABAergic transmission (Ruiz et al., 2004).
Endogenous modulation by kinases/phosphatases
The intracellular loops of β and γ subunits are phosphorylated at a number of
residues (Table 1.2) found within consensus sequences for a number of protein
kinases. Phosphorylation at these sites may increase or decrease GABAergic
currents by modulating receptor gating properties and/or trafficking processes
that determine the number of receptors at the cell surface (see reviews: Smart,
1997; Brandon et al., 2002; Kittler & Moss, 2003; Song & Messing, 2005;
Arancibia-Carcamo & Kittler, 2009; Houston et al., 2009b). The precise effects
on GABAergic currents depend not only on the kinase and subunit isoforms in
question, but also seem to vary with the preparation being studied. For
example, Ca2+/calmodulin-dependent kinase II (CaMKII) failed to modulate
receptors expressed in human embryonic kidney 293 (HEK293) cells, but did so
if receptors were expressed in a neuroblastoma cell line (Houston et al., 2009b).
Protein kinase C (PKC)-dependent phosphorylation can increase cell surface
expression of receptors by disrupting the interaction between β/γ subunits and
AP2 (Kittler et al., 2000; Kittler et al., 2005; Luscher et al., 2011a). Interestingly,
protein kinase B (PKB/Akt)-dependent phosphorylation of the same residues as
PKC, rather than influencing receptor internalisation, increases cell surface
delivery of newly-synthesised receptors – i.e. phosphorylation effects depend
not only on the residues acted upon, but also where in the cell it takes place
(Luscher et al., 2011a). It must not be forgotten that GABAA receptors are not
the only target of these enzymes: Connolly et al. (1999) found that stimulating
PKC in HEK293 cells decreased cell surface expression of α1β2γ2 receptors,
but that the effect does not require phosphorylation at β2S410, γ2S327 or γ2S343.
Chapter 1: Introduction 24
GABAA receptor subunit Residues Kinases
β1
S409 PKA1, PKB
5, PKC
1,3, PKG
1
S384, S409 CaMKII4
β2 S410 PKA1, PKB
2,5, PKC
1,3, PKG
1, CaMKII
1,4
β3
S408, S409 PKA1, PKB
5, PKC
1,3, PKG
1
S383, S409 CaMKII4
γ2
S327, S343 PKC1,3
, CaMKII1,4
S348, T350 CaMKII4
Y365, Y367 Tyrosine kinases1,2
Table 1.2 – Phosphorylation sites on GABAA receptors
GABAA receptor β and γ subunit phosphorylation sites are listed alongside the kinases
that are able to phosphorylate them. Underlined residues are unique to the γ2L isoform
(i.e. absent from the γ2S isoform). Table compiled using reviews by 1Brandon et al.
(2002), 2Luscher and Keller (2004), 3Song and Messing (2005), 4Houston et al. (2009b),
5Luscher et al. (2011a).
Phosphorylation is a readily-reversible post-translational modification, and so
receptor phosphorylation state is likely to be dynamically regulated in vivo.
Indeed, brain-derived neurotrophic factor (BDNF) stimulation of cultured
hippocampal and cortical neurons induces a transient rise and fall in β3 subunit
phosphorylation (Jovanovic et al., 2004; Kanematsu et al., 2006). Synaptic
current amplitudes increase in parallel with the PKC-mediated β3
phosphorylation, which is temporally associated with increased cell surface
expression of receptors (Jovanovic et al., 2004). The latter stages of the BDNF
response, where synaptic current amplitudes diminish and β3 subunits are
dephosphorylated, involve the recruitment of protein phosphatases 1 and 2A
(PP1/PP2A) by PRIP (Kanematsu et al., 2006). It is not clear whether this latter
stage involves reduced cell surface expression (Kanematsu et al., 2006), or
reduced currents without an altered surface expression (Jovanovic et al., 2004).
Nevertheless, these investigations concur that β3 phosphorylation is only
transient in nature after BDNF treatment.
Chapter 1: Introduction 25
Neurosteroids as endogenous modulators
Steroid compounds not only modulate gene expression by binding to
transcription factors of the nuclear receptor family, but they also have non-
genomic actions within the CNS (Paul & Purdy, 1992; Rupprecht & Holsboer,
1999; Belelli et al., 2006). These “neuroactive steroids”, or “neurosteroids”, are
compounds that are produced endogenously and can affect the function of
several neuronal receptors, including NMDA (N-Methyl-D-aspartate), AMPA (2-
amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl)propanoic acid), kainate, glycine,
serotonin, nicotinic acetylcholine, oxytocin, sigma-type-1 and GABAA receptors
(Rupprecht & Holsboer, 1999; Mellon & Griffin, 2002; Strous et al., 2006). The
two major neurosteroids in vivo, allopregnanolone and allo-tetrahydro-deoxy-
corticosterone (THDOC), are among the most potent known endogenous
modulators of GABAA receptors (Paul & Purdy, 1992). Immunohistochemical
analysis demonstrates that these two neurosteroids are mostly concentrated in
cell bodies and dendrites of excitatory neurons, with little or no labelling of
gliaform cells or inhibitory interneurons, suggesting that they act in a paracrine
or autocrine manner to modulate principal cell firing (Saalmann et al., 2007).
That GABAA receptors are a major molecular target for neurosteroids was made
apparent by a combination of radioisotope (measuring 36Cl- flux into
synaptosomes) and electrophysiological studies. These studies demonstrated
that the action of neurosteroids and their analogues is biphasic: low (nM)
concentrations augment GABA-induced Cl- conductance, whereas higher (µM)
concentrations directly stimulate GABAA receptor activation – producing slow
inward currents that resemble the direct responses to pentobarbital (Harrison &
Simmonds, 1984; Harrison et al., 1987; Puia et al., 1990; Paul & Purdy, 1992).
The majority of estimates for in vivo neurosteroid concentrations support a role
for potentiation at GABAA receptors, but sufficient levels to directly activate
these receptors may be achieved under specific circumstances, such as in
mother and foetus during late pregnancy (Ichikawa et al., 1974; Paul & Purdy,
1992; Biedermann & Schoch, 1995; Luisi et al., 2000; Nguyen et al., 2003).
Chapter 1: Introduction 26
At the molecular level, neurosteroids are thought to act by altering the kinetics
of receptor entry into and exit from desensitised states (Zhu & Vicini, 1997) and
may increase the efficacy of ion channel gating (Bianchi & Macdonald, 2003). At
the cellular level, neurosteroid-mediated potentiation at GABAA receptors
enhances synaptic (Belelli & Herd, 2003; Harney et al., 2003) and tonic currents
(Stell et al., 2003). Both responses are expected to hyperpolarize the cell
membrane and/or shunt excitatory inputs, and thus reduce the neuron’s
probability of firing. There is also a class of neurosteroids that antagonize
GABAA function: the so-called “excitatory steroids” – mostly sulphated forms of
the classical neurosteroids, such as pregnenolone sulphate, or 3β-hydroxy-
steroids (Akk et al., 2001; Wang et al., 2002). They are pro-convulsant in
animals, act as non-competitive antagonists at the GABAA receptor, and
probably bind to a site distinct from potentiating neurosteroids; their effects in
animals may also be due to an enhancement of excitatory glutamatergic
transmission (Paul & Purdy, 1992; Akk et al., 2001; Wang et al., 2002; Hosie et
al., 2007).
In our study, we hope to increase understanding of the roles played by
endogenous neurosteroids, specifically via positive allosteric modulation at α2-
type GABAA receptors. We have particularly focused on their roles in the aetio-
pathology of anxiety and depression. Furthermore, we have explored
therapeutic potential for α2-subunit-targeting neurosteroids in these disorders.
The following sections therefore provide details regarding current knowledge of
neurosteroid physiology (Section 1.2) and its links to anxiety (Section 1.3) and
depression (Section 1.4).
Chapter 1: Introduction 27
1.2. Neurosteroid physiology and pharmacology
1.2.1. Endogenous neurosteroids: synthesis and roles
Neurosteroids are synthesized in vivo from a cholesterol precursor, via a range
of enzymatic conversions (see Fig. 1.3). Neurosteroid distribution within the
brain is not uniform, with the highest levels generally in the olfactory bulb,
striatum and cortex, and lower levels in brainstem, suggesting there will be a
regional variation in neurosteroidogenesis and neurosteroid-mediated
modulation of GABAergic transmission (Uzunov et al., 1996; Bixo et al., 1997;
Bernardi et al., 1998; Saalmann et al., 2007). Neurosteroids and their
precursors can also be produced in peripheral steroidogenic tissues such as the
adrenal cortex and gonads. For example, menstrual-cycle-related changes in
circulating progesterone can influence brain allopregnanolone levels, with key
consequences for sufferers of catamenial epilepsy (see Section 1.2.4).
Furthermore, stress-induced rises in brain neurosteroid levels are thought to
involve production of these compounds in the adrenal glands (Purdy et al.,
1991; Barbaccia et al., 1998). However, a change in peripheral levels cannot
necessarily be projected to cause changes in CNS levels of a given
neurosteroid. Indeed, the rises in serum and brain allopregnanolone and
THDOC levels during pregnancy do not directly correlate with the rise in serum
progesterone (in rats (Concas et al., 1998) or humans (Luisi et al., 2000)) – i.e.
peripheral progesterone is not the only source for brain allopregnanolone.
Chapter 1: Introduction 28
Figure 1.3 – Neurosteroid synthesis pathways
The steps involved in converting cholesterol into GABAA receptor modulating
neurosteroids are shown. A number of additional neuroactive compounds can also be
produced by modifications of the various intermediates (black) (for details, see Mellon
& Griffin, 2002; Stoffel-Wagner, 2003), but this figure focuses on the major positive
(green) and negative (red) allosteric modulators of GABAA receptors. The neurosteroid
profile of a particular brain region will depend on the enzymes present locally; in
humans, hippocampal and temporal lobe expression and/or enzymatic activity has
been demonstrated for P450scc, 21β-hydroxylase, 5α-reductase and 3α-HSD (Stoffel-
Wagner, 2003). The initial conversion of cholesterol to pregnenolone occurs within
mitochondria, requiring the activity of two transporters: StAR (steroidogenic acute
regulatory protein (Stocco & Clark, 1996)) and TSPO (the 18 kDa translocator protein),
the latter of which is sensitive to stimulation by benzodiazepines (Papadopoulos &
Lecanu, 2009). Abbreviations: 3β-HSD, 3β-hydroxy-steroid dehydrogenase; 3α-HSD,
3α-hydroxy-steroid dehydrogenase; P450scc, cytochrome P450 cholesterol side-chain
cleavage; P450c17, 17α hydroxylase, c17,20 lyase. Note that the HSD enzymes are
also referred to as hydroxy-steroid oxido-reductase (HSOR) enzymes.
The levels of neurosteroids have been measured most commonly by radio-
immunoassay (Purdy et al., 1990a) or by mass fragmentography (MF) (Uzunov
et al., 1996). The former method is limited by the antibody specificity, but pre-
separation of the various steroids extracted from the sample by high-
Chapter 1: Introduction 29
performance liquid chromatography (HPLC) can eliminate the problem of cross-
reactivity with other neurosteroid metabolites (Purdy et al., 1990a). The MF
approach is more technically demanding, but has been proposed to be a more
sensitive approach, capable of accurately measuring picomolar levels of steroid
(Uzunov et al., 1996). The estimates of baseline rodent brain neurosteroid
levels vary, with reports for allopregnanolone ranging from very low (<3 nM
(Purdy et al., 1991; Vallee et al., 2000)), to levels sufficient to potentiate GABAA
receptors (3-10 nM (Uzunov et al., 1996); 2-20 nM (Bernardi et al., 1998)).
Baseline allopregnanolone levels in human brains may be higher (e.g. in
women, may range from 30-70 nM, depending on serum progesterone levels
(Bixo et al., 1997)), but human brain samples are less readily available, and
obtained post-mortem.
Allopregnanolone and THDOC levels increase during pregnancy (Concas et al.,
1998; Luisi et al., 2000) and in response to stressors (Purdy et al., 1991; Paul &
Purdy, 1992) or drugs – including nicotine (Porcu et al., 2003), gamma-hydroxy
butyrate (Barbaccia et al., 2002) and some antidepressants (Uzunov et al.,
1996). Increased production of endogenous neurosteroids is also believed to
mediate some of the behavioural responses to ethanol, including its
antidepressant and anticonvulsant effects (VanDoren et al., 2000; Hirani et al.,
2002; Khisti et al., 2002; Helms et al., 2012).
Dysregulation of neurosteroids is associated with a range of diseases, including
premenstrual dysphoric disorder, panic disorder, anxiety and stress,
depression, schizophrenia and bipolar disorder, eating disorders, and dementia
(Mellon & Griffin, 2002; Strous et al., 2006). It is often difficult to determine
whether the altered neurosteroid levels seen in some of these diseases is a
consequence or cause of the condition, especially because some steroids have
biphasic effects (e.g. pregnenolone sulphate is anxiolytic at low doses, but
anxiogenic at higher doses (Strous et al., 2006)). The roles for neurosteroids in
anxiety and depression will be discussed in more detail below (Sections 1.3 and
1.4).
Chapter 1: Introduction 30
1.2.2. Neurosteroid binding to GABAA receptors
The effects of a range of neurosteroids and their analogues on GABAA receptor
activity demonstrated stereospecificity and a biphasic action (potentiation at low
concentration, activation at high concentration (Purdy et al., 1990b; Paul &
Purdy, 1992)). It was therefore proposed that there would be two specific
neurosteroid binding sites on the GABAA receptor: an “activation site” and a
“potentiation site”. Identification of these sites was not a trivial task (see review
by Hosie et al., 2007), but was eventually achieved by electrophysiological
characterisation of site-directed mutants expressed in HEK293 cells (Hosie et
al., 2006); the key residues for neurosteroid function are highlighted on Fig. 1.4.
Note that with this updated GABAA receptor model, the activation site residues
identified by Hosie et al. (2006), now point away from the interface between α
and β subunits, where neurosteroid was proposed to bind. Indeed, more recent
etomidate photolabelling suggests a model of the transmembrane domain (Li et
al., 2009; Olsen & Li, 2011) that is incompatible with that of Hosie et al. (2006).
Residues βY284 and αT236 are therefore more likely to be involved in a
transduction mechanism, rather than direct binding of neurosteroid. The
orientations of residues in the potentiation site, on the other hand, are still
compatible with the model of Hosie et al. (2006).
When heterologously-expressed receptors are studied, there are small isoform-
dependent differences in responses that may be of relevance when considering
the low nM concentrations of neurosteroids found in vivo: minimal potentiating
doses for allopregnanolone were 3 nM at α1β1γ2 and α3β1γ2, 10 nM at
α6β1γ2, and 30 nM at α2/4/5β1γ2 subunit combinations (Belelli et al., 2002).
Interestingly, receptors incorporating a δ subunit show increased efficacy
relative to equivalent γ-containing combinations (Belelli et al., 2002; Bianchi &
Macdonald, 2003; Lambert et al., 2003; Hosie et al., 2009). However, this effect
is probably not due to an interaction of neurosteroid with the δ subunit, because
the potentiation site is entirely confined within the α subunit (Hosie et al., 2009).
Furthermore, a single point mutation of the conserved potentiation-site
glutamine in the α4 subunit to leucine (α4Q246L) ablates neurosteroid potentiation
Chapter 1: Introduction 31
of α4β3δ receptors (Hosie et al., 2009). The increased sensitivity of αβδ
combinations is probably a reflection of the partial agonist nature of GABA at
these receptors (c.f. αβγ receptors, where GABA is a full agonist) (Bianchi &
Macdonald, 2003).
Figure 1.4 – GABAA receptor subunit assembly and neurosteroid binding sites
Ribbon diagrams depict the secondary structure of the transmembrane regions of α
(green) and β (red) subunits, viewed in the plane of the membrane, and detail the
proposed neurosteroid potentiation (A) and direct activation (B) sites. Structures are a
3D homology model of an α1β2γ2 GABAA receptor, based on the crystal structure of
the Caenorhabditis elegans glutamate-gated chloride channel, GluCl (Hibbs & Gouaux,
2011), courtesy of Dr. Marc Gielen. Residues important for neurosteroid binding are
highlighted as solid spheres, whilst the proposed neurosteroid binding pockets (Hosie
et al., 2006) are indicated by grey ovals. The potentiation site is contained within the α
subunit, involving transmembrane helices M1 and M4 (key residues in α1 are Q241,
N407 and Y410, which are conserved across all α subunit isoforms). Within this model,
the position of key conserved residues in the direct activation site – α1 T236 and β2
Y284 – appear inconsistent with the proposed neurosteroid binding roles for these
residues.
Chapter 1: Introduction 32
1.2.3. Physiological modulation of GABAA receptors by neurosteroids
The neurosteroids allopregnanolone and THDOC enhance GABAergic
transmission by slowing the decay of IPSCs (Belelli & Herd, 2003; Harney et al.,
2003) and/or increasing the magnitude of tonic currents (Stell et al., 2003).
Some investigators have proposed that physiological (low nM) levels of
neurosteroid will be ineffective at synaptic GABAA receptors – for example, Stell
et al. (2003) found that DG GC tonic currents were sensitive to 10 nM THDOC,
but phasic events only to 100 nM of this neurosteroid. Because tonic currents
are often carried by αnβnδ combinations (e.g. in DG GCs (Nusser & Mody,
2002; Stell et al., 2003) and cerebellar GCs (Brickley et al., 2001)), they might
be better poised to respond to endogenous steroids than synaptic αnβnγ
combinations (Bianchi & Macdonald, 2003). However, there is some evidence
for modulation of phasic events by physiological neurosteroid levels. Firstly,
Puia et al. (2003) showed that decreasing endogenous allopregnanolone levels
in neocortical brain slice tissue (using SKF-10511, an inhibitor of 5α-reductase
type I and II) speeds the IPSC decay times recorded from pyramidal neurons.
Furthermore, CA1 PC miniature IPSCs (mIPSCs) respond to 10 nM
allopregnanolone (Harney et al., 2003), whilst tonic currents in these cells
appear neurosteroid-insensitive (Stell et al., 2003). It therefore seems likely that,
in vivo, endogenous neurosteroids will modulate both phasic and tonic currents,
with the extent of modulation of each current depending on the neurosteroid-
sensitivity of the underlying GABAA receptor composition.
Interestingly, the sensitivities of various neurons to neurosteroids do not always
correlate with the relative sensitivities of the major GABAA receptor isoforms
expressed in these cells (Belelli et al., 2006). By a comparison of responses
across the literature, it would seem that IPSC sensitivity to neurosteroid can
vary with cell type (Harney et al., 2003) and age of animal (Cooper et al., 1999;
Mtchedlishvili et al., 2003), and may depend on differences in local neurosteroid
metabolism (Belelli & Herd, 2003), receptor subunit composition (Brussaard et
al., 1997; Cooper et al., 1999; Mtchedlishvili et al., 2003) and relative kinase
and phosphatase activities (Brussaard et al., 2000; Fancsik et al., 2000; Harney
Chapter 1: Introduction 33
et al., 2003; Koksma et al., 2003). It is possible that some cells would respond
to the basal in vivo levels of neurosteroids, whilst others will only be modulated
by neurosteroids when concentrations are increased – by stress or pregnancy,
for example (e.g. Harney et al. (2003) show that synaptic events in CA1 PCs
respond to basal allopregnanolone levels (10 nM), whilst DG GCs will only
respond to heightened levels (300 nM)).
As was discussed in Section 1.1.3, the effects of phosphorylation on GABAA
receptor function appear to depend on the receptor subunit composition, the
kinase, and the cell-type studied. These observations can be extended to
kinase/phosphatase modulation of neurosteroid sensitivity. For example,
Harney et al. (2003) attribute the differential allopregnanolone sensitivity
(measured by IPSC decay prolongation) of hippocampal PCs and DG GCs to
distinct activities of PKC and PKA within these cells. Constitutive activity of PKC
and G protein appears to be required for allopregnanolone sensitivity in
magnocellar neurons of the supraoptic nucleus (SON) of rats: IPSC decay
prolongation by 1 µM of this neurosteroid is prevented by inhibitors of these
proteins, whilst activators of these proteins do not further enhance the response
(Fancsik et al., 2000). Interestingly, the latter cell type is unaffected by inhibition
of PKA, further supporting the notion that there is a cell-type variation in the
kinases that affect neurosteroid modulation (i.e. PKA activity is more important
in CA1 PCs (Harney et al., 2003) than SON magnocellar neurons (Fancsik et
al., 2000)). Even within the same cell type, however, there is conflicting
evidence for the effects of phosphorylation: unlike the work by Fancsik et al.
(2000), others find that inhibition of PKC is required for neurosteroid sensitivity
of SON magnocellar neurons and that constitutive activity of phosphatases
PP1/PP2A determine allopregnanolone sensitivity of these cells during
pregnancy (Brussaard et al., 2000; Koksma et al., 2003). Some of the
discrepancies in the literature may relate to the use of broad-range kinase
modulators (e.g. phorbol esters stimulate a range of PKC isoforms, which may
have opposing effects on receptor function (Song & Messing, 2005)).
GABAergic currents in vivo are probably modulated in concert by
kinases/phosphatases and neurosteroids. For example, a combination of falling
neurosteroid levels, and reduced neurosteroid sensitivity (by GABAA receptor
Chapter 1: Introduction 34
phosphorylation), are thought to contribute to the increased neuronal firing
required for timed release of oxytocin at parturition (Brussaard et al., 2000;
Koksma et al., 2003).
1.2.4. Neurosteroids: therapeutic potential
Historically, synthetic neurosteroid analogues were used for general
anaesthesia during surgery (Prys-Roberts & Sear, 1980). Although the first
evidence of anaesthetic action for neurosteroids came in the 1940s for
progesterone (Paul & Purdy, 1992; Belelli et al., 2006), it was not until the
1980s that this response was shown to involve an enhancement of GABAA
receptor function (Harrison & Simmonds, 1984; Harrison et al., 1987; Paul &
Purdy, 1992). Despite their favourable potencies and safety profiles, these
drugs are no longer used in humans due to the frequency of anaphylactic
reactions, although these may be due to the Cremphor EL vehicle used (Prys-
Roberts & Sear, 1980). There is resurging interest in translating neurosteroid-
based therapies to the clinic because of their potency and array of favourable
effects when studied in animal models: neurosteroids can be anxiolytic,
antidepressant, analgesic, sedative, anticonvulsant, anaesthetic, and can
suppress the action of the stress-responsive hypothalamic-pituitary-adrenal
(HPA) axis (Carter et al., 1997; Barbaccia et al., 1998; Rodgers & Johnson,
1998; Khisti et al., 2000; Winter et al., 2003; Belelli et al., 2006).
The clinical potential for neurosteroids has been well supported for treatment of
Niemann Pick type C disease and epilepsy. The former is an inherited
lysosomal storage disorder where patients suffer rapid neurodegeneration,
which is ultimately fatal. Mouse models of Niemann Pick disease have very low
allopregnanolone levels, and the administration of allopregnanolone delays the
onset of motor dysfunction and extends survival (i.e. modulation of GABAA
receptors by this neurosteroid is crucial for proper brain development and
survival (Griffin et al., 2004; Mellon et al., 2008)). The anti-seizure activity of the
Chapter 1: Introduction 35
neurosteroid analogue, ganaxolone, has been extensively verified in animal
models of epilepsy (Carter et al., 1997; Reddy & Rogawski, 2009, 2010) and it
has had some success in human clinical trials (Nohria & Giller, 2007). There is
also anecdotal evidence favouring the use of ganaxolone in catamenial epilepsy
(Reddy & Rogawski, 2009). In this subtype of epilepsy, sufferers experience an
increase in seizure frequency and severity in response to an ovarian-cycle-
linked fall in allopregnanolone levels, and an associated increase in expression
of α4-type GABAA receptor subunits (Smith et al., 1998; Reddy & Rogawski,
2009).
To establish and appreciate the full therapeutic potential of neurosteroids and
their analogues, it is also important to fully define their normal physiological
actions at molecular, cellular and systems levels. We predict that neurosteroid
functions will be separable according to α subunit isoform (as was determined
for benzodiazepines: see Table 1.1). Our intention was therefore to explore the
physiological functions and therapeutic potential of neurosteroids in anxiety and
depression. The following sections will outline the current understanding in
these fields.
1.3. Neurosteroids and GABAA receptors in anxiety
1.3.1. The HPA axis and neurosteroids as endogenous anxiolytics
The HPA axis comprises a set of interactions between the hypothalamus,
pituitary gland and adrenal cortex, and is activated by stress (Mody & Maguire,
2011). Activation of the axis stimulates production of a series of steroid
hormones in the adrenal cortex, including allopregnanolone and THDOC.
Activation of the rat HPA axis by exposure to stress, such as foot-shock or
forced swim, or by inhibiting GABA synthesis or negatively modulating GABAA
Chapter 1: Introduction 36
receptors, therefore induces an increase in brain and plasma neurosteroid
levels (Purdy et al., 1991; Barbaccia et al., 1996; Barbaccia et al., 1998; Vallee
et al., 2000). These rises in neurosteroid are disrupted by adrenalectomy and
castration, indicating that increased production of these compounds occurs
mostly in the periphery, rather than in the CNS (Purdy et al., 1991; Barbaccia et
al., 1998). Indeed, after this surgery, basal and stress-associated levels of
THDOC are undetectable (Purdy et al., 1991). Basal allopregnanolone levels
are also greatly lowered after adrenalectomy, but this steroid is still detectable
in the cerebral cortex after such treatment, indicating another source for its
production (probably the CNS) (Purdy et al., 1991). By examining the time-
courses of behavioural and neurochemical responses to acute stress (Purdy et
al., 1991; Barbaccia et al., 1998), this HPA-induced rise in neurosteroids had
been proposed to represent a feed-back mechanism to recover normal
GABAergic tone and HPA function after acute stress. Given that injected
neurosteroids reduce anxiety (Crawley et al., 1986; Wieland et al., 1991), the
stress-induced production of neurosteroids is thought not only to limit the
neurochemical response to stress, but also the behavioural response – i.e.
neurosteroids probably function as endogenous anxiolytics (Barbaccia et al.,
1998).
The hippocampus is one of the main inputs to the hypothalamus that triggers
HPA axis activation in stress, and is probably a key site for this neurosteroid-
mediated feedback (by enhancing inhibitory neurotransmission in the
hippocampus). Excessive activation of the HPA axis by chronic stress can be
particularly damaging to the hippocampus, and reduces the number of
parvalbumin-positive interneurons (Hu et al., 2010). Diminished inhibitory
neurotransmission within the hippocampus could therefore trigger pathological
positive feedback – further activating the HPA axis (with glucocorticoids further
damaging the hippocampus (Brown et al., 1999; Hu et al., 2010)). Such HPA
axis dysregulation may account for the hippocampal atrophy observed in
patients suffering from post-traumatic stress disorder and depression (Brown et
al., 1999; Sheline, 2003). Furthermore, the consequent impairment of regulation
of network oscillations may underlie the cognitive defects in stress-related
disorders (Hu et al., 2010). We have therefore focussed part of our investigation
Chapter 1: Introduction 37
on inhibitory neurotransmission in the hippocampus, probing whether
neurosteroid modulation of this transmission involves the α2-type GABAA
receptor.
1.3.2. Defining the GABAA receptor α subunits that are important in anxiety and
anxiolysis
An assortment of evidence is available to support a role for α2-type GABAA
receptors in anxiety and anxiolysis. Firstly, this subunit is enriched in brain
regions linked to emotion and anxiety, including the hippocampus and
amygdala (Sperk et al., 1997; Pirker et al., 2000; Sieghart & Sperk, 2002;
Mohler, 2006a). Furthermore, α2 knock-out (α2-/-) mice are anxious in a
conditioned emotional response paradigm (Dixon et al., 2008) and α2H101R
knock-in mice (benzodiazepine-insensitive at α2-type GABAA receptors) lose
the anxiolytic response to diazepam (Low et al., 2000).
Not all investigators agree, however, that α2 subunits are the sole mediators of
anxiolysis. In an investigation by Low et al. (2000), mice insensitive to
benzodiazepine action at the α3 subunit retain anxiolysis following diazepam
administration, and so investigators concluded that this subunit was not
involved in anxiety. This notion is further supported by observations that α3-/-
mice are not anxious and retain diazepam anxiolysis (Yee et al., 2005). On the
other hand, an α3-selective inverse agonist is anxiogenic (Atack et al., 2005)
and an α3-selective agonist is anxiolytic (Dias et al., 2005), both of which
implicate α3 subunits in anxiety circuitry. Notably, there are confounding issues,
such as problems with activity effects in locomotor-dependent anxiety tests (see
discussion by Reynolds et al., 2001), or issues of imperfect selectivity of
subunit-selective compounds (e.g. the α3-selective inverse agonist shows some
efficacy at α2 subunits (Atack et al., 2005), and so its anxiogenesis could be α2-
mediated). Nevertheless, both α2 and α3 subunits mediate the myorelaxant
effects of benzodiazepines: α2 subunits at low doses, and α3 subunits at higher
Chapter 1: Introduction 38
doses (Crestani et al., 2001). It would seem probable that there is an
overlapping contribution of both subunits in anxiolysis.
It is generally accepted that α1 subunits mediate the sedative effects of
benzodiazepines (Rudolph et al., 1999). Investigators therefore hoped that
creating a α2/α3-selective benzodiazepine-like compound that has little activity
at α1 subunits would be expected to produce anxiolysis without sedation.
Several such compounds have been generated, including TPA023 (also called
MK-0777) and MRK-409 (also called MK-0343), which have proven successful
as non-sedating anxiolytics in pre-clinical assessments (Atack, 2008; Atack et
al., 2011). However, in spite of promising pre-clinical results, subunit-selective
compounds have yet to be translated to the clinic. TPA023 failed at phase II
clinical trials due to toxicity in long-term dosing (Mohler, 2011). MRK-409
unexpectedly proved sedative in humans, despite a lack of sedation in
preclinical models; investigators believe this effect is a consequence of partial
agonist efficacy at α1 subunits, but indicates that humans have a greater
sedation tendency than animal models (Atack, 2010; Atack et al., 2011).
Confusingly, Ocinaplon is a compound that shows an anxioselective effect in
humans and animal models, but has greater efficacy (both absolute and relative
to diazepam) at α1- than α2-type GABAA receptors (Lippa et al., 2005).
Nevertheless, the compound TPA023 at least provides proof of principle that
α2-selective/α1-sparing compounds can be non-sedating anxiolytics in human
subjects (Atack, 2008).
We propose that neurosteroids will exhibit a similar α-subunit-selective profile,
and thus predict a role for α2-type GABAA receptors in anxiolytic responses to
both endogenous and injected neurosteroids. Furthermore, if this hypothesis is
correct, α2-subunit-selective neurosteroid analogues would be non-sedating
anxiolytics – opening an alternative drug design avenue to the selective
benzodiazepine approach.
Chapter 1: Introduction 39
1.4. Neurosteroids and GABAA receptors in depression
1.4.1. Roles for GABAA receptors in depression
For many years, a prevailing theory for depression was the ‘monoamine
hypothesis’, which postulates that depression results from a reduction in
monoamine neurotransmission. Iproniazid and imipramine had demonstrated
antidepressant efficacy in humans, and were later shown to increase serotonin
and noradrenaline transmission – suggesting that deficiencies in these
neurotransmitters underlie depression (see review by Krishnan & Nestler,
2008). Many antidepressant agents therefore aim to raise synaptic levels of
monoamines, either by blocking degradation (e.g. tranylcypromine, which
inhibits monoamine oxidases) or re-uptake into neurons (e.g. selective
serotonin reuptake inhibitors (SSRIs)). Despite their widespread use, these
drugs are not universally effective, and it is increasingly appreciated that the
monoamine hypothesis has some shortcomings (Lacasse & Leo, 2005).
Alternatives to the monoamine hypothesis are emerging, with additional modes
of action being proposed for common antidepressants, as well as new
suggestions for the aetiopathology of depression, including a role for
GABAergic signalling.
Luscher et al. (2011b) and Smith and Rudolph (2012) reviewed the literature in
support of a role for GABAergic dysfunction in depression. This connection
might not be a surprise, given that anxiety and depression are often co-morbid
(Hirschfeld, 2001; Nutt et al., 2006), and that GABAA receptors are linked with
anxiety (see Section 1.3.2). Co-morbid anxiety and depression is a significant
challenge because patients often present with greater symptom severity and
have delayed or diminished response to conventional treatment (Hirschfeld,
2001; Nutt et al., 2006). A better understanding of the common mechanisms for
anxiety and depression may therefore help develop better treatment strategies
for such patients.
Chapter 1: Introduction 40
Arguably the best demonstration of a causative role for GABAergic deficits in
anxiety and depression comes from study of heterozygous mouse knock-outs
for the γ2 subunit of GABAA receptors (γ2+/-), which display phenotypes
consistent with both human disorders, including increased avoidance of
aversive environments, increased despair and enhanced sensitivity to
ambiguous cues (Crestani et al., 1999; Earnheart et al., 2007). Furthermore,
this mouse model recapitulates the HPA axis hyperactivity seen in melancholic
forms of depression (Shen et al., 2010). Particularly important for the purposes
of this study, however, are observations that α2-/- mice have phenotypes
consistent with both anxiety and depression (Dixon et al., 2008; Vollenweider et
al., 2011) and that several limbic regions whose structure and/or activity is
altered in depression, including the hippocampus, amygdala and basal ganglia
(Sheline, 2003; McCabe et al., 2009), are areas rich in GABAA receptor α2
subunit expression (Wisden et al., 1992; Sperk et al., 1997; Pirker et al., 2000;
Sieghart & Sperk, 2002; Mohler, 2006a).
There is also evidence from humans in support of a GABAergic signalling deficit
in depression. A correlation between depressive symptoms and reduced brain
levels of GABA was noted in cortical biopsies taken from fourteen depressed
patients (Honig et al., 1988). Although not necessarily representative of brain
levels, plasma GABA concentrations can be more conveniently determined, and
were also diminished in approximately 40% of depressed patients compared to
healthy controls (Petty, 1994). Non-invasive means of measuring brain GABA
levels were later developed, such as the proton magnetic resonance
spectroscopic approach of Sanacora et al. (1999). Their work confirmed a
GABA deficit in the occipital cortex (Sanacora et al., 1999), and prefrontal
regions (Hasler et al., 2007) of depressed individuals, the latter of which is more
likely to be of significance for mood disorders.
A role for GABAA receptors in human depression is supported by examining
subunit mRNAs in post-mortem samples, mostly from depressed suicide
victims. These studies found increases and decreases in expression of various
Chapter 1: Introduction 41
subunits, including α1, α3, α4 and α5 (Luscher et al., 2011b). However, it is
important to note that these mRNA fluctuations may not necessarily be
paralleled by cell surface protein expression levels. By employing single photon
emission computed tomography (SPECT) with [123I]iomazenil to estimate the
number of cortical GABAA receptors, Kugaya et al. (2003) found no change
between 13 depressed patients and 19 healthy controls, which may negate the
above findings with mRNA levels. Nevertheless, the SPECT approach is not
without its limitations, particularly an inability to determine which subunit
combinations make up the GABAA receptors to which [123I]iomazenil is bound.
Kugaya et al. (2003) therefore cannot rule out an upregulation of one subunit at
the expense of another, despite no change in the overall number of
[123I]iomazenil binding sites. By using post-mortem slices of hippocampus from
Bipolar 1 patients and healthy controls, Dean et al. (2005) were able to concur
with such a mechanism: overall binding of [3H]flumazenil was unaltered (i.e. no
change in overall GABAA receptor expression), but the proportion of
[3H]flumazenil binding that was sensitive to displacement by zolpidem was
lowered (note that, unlike [3H]flumazenil, zolpidem does not bind to α5
subunits). The interpretation of these observations is therefore that Bipolar 1
patients have an increased expression of α5 subunits, at the expense of other α
subunits in their hippocampi.
Several genetic association studies fail to link GABAA receptor α subunits with
major depression and anxiety disorders (Henkel et al., 2004; Pham et al., 2009).
However, these studies used a broad range of subjects with anxiety and
depressive symptoms, and links may have been diluted by such
generalisations; for example, Henkel et al. (2004) do find an association
between GABAA α3 subunit and unipolar major depressive disorder specifically
in female subjects. Furthermore, Yamada et al. (2003) find association between
polymorphisms on genes encoding GABAA receptor α1 and α6 subunits and
mood disorders in female patients. Horiuchi et al. (2004) also link an α1 subunit
polymorphism with a cohort of Japanese patients suffering affective disorders.
Chapter 1: Introduction 42
Reductions in brain tissue volume in several areas in depression (reviewed by
Sheline, 2003) could represent excitotoxic loss of neurons and/or a reduction in
neurogenesis. Support for reduced neurogenesis comes from the γ2+/- mouse
model of depression, where reduced γ2 expression impairs survival or
differentiation of neuronal precursors into mature neurons (Earnheart et al.,
2007). Interestingly, the effects of reducing γ2 levels on hippocampal
neurogenesis and animal behaviour depend on the developmental stage at
which the deficit is imposed. A conditional knock-out of γ2 after the fourth
postnatal week fails to recapitulate effects of losing this subunit during
embryonic development (Earnheart et al., 2007; Shen et al., 2010). The story is
not complete, however, since the developmentally-delayed loss of γ2 subunits
in the model described above retains the HPA axis hyperactivity characteristic
of depression (Shen et al., 2010), but fails to develop the hippocampal
neurogenesis defects and behavioural depression, that occurs if the γ2 deficit
is imposed during embryogenesis (Earnheart et al., 2007).
Given the above indications that depression involves reduced transmission
through GABAA receptor subunits, one might predict that potentiating this
transmission would be a successful treatment strategy. Indeed, Luscher et al.
(2011b) suggest that current antidepressants, particularly SSRIs, may
potentiate GABAergic transmission in a number of ways: by increasing
excitability of GABAergic interneurons; by increasing GABA production via the
enzyme, glutamate decarboxylase 67 (GAD67); and by increasing levels of
GABAA receptor-potentiating neurosteroids (see Section 1.4.2). However,
benzodiazepines – classical potentiators of GABAergic transmission – are not
generally antidepressant; only alprazolam has demonstrated any efficacy in
depression, although its use is mostly avoided due to its dependence liability
(Laakmann et al., 1996). More work clearly needs to be done to fully understand
the roles GABAA receptor dysfunction in depression. We aim to contribute to
this by further assessing the roles played by α2-type GABAA receptors in
depression.
Chapter 1: Introduction 43
1.4.2. Roles for neurosteroids as antidepressants
Plasma and cerebrospinal fluid (CSF) levels of the neurosteroid
allopregnanolone have repeatedly been shown to be reduced in depressed
patients relative to healthy controls (Romeo et al., 1998; Uzunova et al., 1998;
Strohle et al., 1999; Strohle et al., 2000). Furthermore, this allopregnanolone
deficit is ameliorated by treatment with a range of antidepressants, with the
recovery of CSF levels (Uzunova et al., 1998), but not plasma levels (Romeo et
al., 1998), correlating with the degree of symptomatic resolution. By
characterising the effects of SSRIs on rat brain neurosteroid profiles (Uzunov et
al., 1996) and on the kinetics of recombinantly-expressed enzymes (Griffin &
Mellon, 1999), the increased allopregnanolone concentration appears to result
from stimulation of the enzyme 3α-HSD (see Fig. 1.3) changing its properties to
favour the reductive reaction over the oxidative reaction. It would therefore
seem that neurosteroids could be used as antidepressants. Importantly,
however, such correlative observations do not demonstrate whether
neurosteroid level normalisation is a consequence or cause of diminished
depressive symptoms. Support for a causative role is provided by observations
that sex hormones and their neurosteroid metabolites modulate behaviour in
several rodent models of depression, including forced swim, tail suspension and
learned helplessness tests (Khisti et al., 2000; Hirani et al., 2002; Dhir &
Kulkarni, 2008; Shirayama et al., 2011). Antagonism of allopregnanolone’s
antidepressant effect by bicuculline implicates GABAA receptors as the
mediators of the response to neurosteroids (Khisti et al., 2000).
At childbirth, the sudden withdrawal from pregnancy-associated high levels of
progesterone, and its neurosteroid metabolites, is thought to trigger post-partum
depression in susceptible women (Bloch et al., 2000). Support for these
proposals comes from progesterone-withdrawal mouse models of post-partum
depression, where repeated progesterone injections are followed by either a
progesterone receptor antagonist or finasteride (an inhibitor of 5α-reductase) to
inhibit further metabolism to neurosteroids. The former demonstrates that some
of the depressed phenotype involves the progesterone receptor (Beckley et al.,
Chapter 1: Introduction 44
2011), but the latter confirms that reduced levels of its metabolites, especially
allopregnanolone, also contribute to the depressed phenotype in the forced
swim test (Beckley & Finn, 2007). Mouse models also support a role of GABAA
receptors in postpartum depression: δ+/- and δ-/- mice are depressed in various
behavioural tests, but only post-partum – virgin females show no phenotype –
which suggests a protective role of δ subunit expression (Maguire & Mody,
2008).
Whilst GABAA receptor potentiation by neurosteroids may account for some
antidepressant-like effects, there are some important points to note. SSRI
antidepressants are slow to act in humans, requiring several weeks of dosing
before a beneficial effect is observed, whilst their effects in rodent behavioural
models of depression are much more immediate, occurring within minutes
(Cryan et al., 2005). In addition, although these drugs rapidly alter 3α-HSD
activity and brain neurosteroid profiles in rats (Uzunov et al., 1996; Griffin &
Mellon, 1999), chronic fluoxetine treatment in rats actually decreases the
plasma and brain concentrations of THDOC and allopregnanolone (Serra et al.,
2002). Furthermore, other antidepressants, such as imipramine, do not
influence 3α-HSD activity or neurosteroid levels in rodents (Uzunov et al., 1996;
Griffin & Mellon, 1999). Finally, sulphated steroids, which negatively modulate
GABAA receptors, are also antidepressant in mice (Dhir & Kulkarni, 2008).
Therefore, although neurosteroid potentiation at GABAA receptors can produce
antidepressant-like effects in animal models, they are by no means a universal
feature of antidepressant function. We therefore aim to screen the
antidepressant potency of the neurosteroid THDOC, and assess whether this
function depends on positive allosteric modulation of α2-type GABAA receptors.
1.4.3. The nucleus accumbens in reward and depression
The nucleus accumbens (NAcc) forms part of the mesolimbic dopamine
pathway, which is believed to process rewarding stimuli, and to be involved in
Chapter 1: Introduction 45
motivational reward-seeking pathways. Major inputs to this structure are
dopaminergic projections from the ventral tegmental area (VTA), glutamatergic
projections from the hippocampus, as well as inputs from prefrontal association
cortices and the basolateral amygdala (Heimer et al., 1997). Drugs of abuse,
including cocaine and benzodiazepines, engender rewarding effects by
increasing dopamine release in the NAcc, often by relieving inhibitory restriction
on VTA dopamine neuron firing (Hyman & Malenka, 2001; Luscher & Ungless,
2006; Tan et al., 2011). NAcc activity is under extensive inhibitory control: the
vast majority (95%) of neurons in the NAcc are GABAergic medium spiny
neurons (MSNs), which not only project to outputs (such as the ventral
pallidum), but also make collateral connections with one another (Heimer et al.,
1997). Changes in GABAergic transmission may occur during addiction and
withdrawal – for example, the level of extracellular GABA increases in the rat
NAcc after withdrawal from daily cocaine injections (Xi et al., 2003).
The mesolimbic dopamine pathway is not only involved in response to addictive
drugs, but is also activated by a number of natural rewards, such as food.
Anhedonia, a reduced pleasure in response to normally rewarding stimuli, is a
core symptom of depression, suggesting that a defective mesolimbic dopamine
pathway can underlie some features of depression (Shirayama & Chaki, 2006).
It is possible that the antidepressant function of neurosteroids could involve
their activity within the mesolimbic dopamine system. Direct injection of
allopregnanolone into the NAcc was not antidepressant in a learned
helplessness model of depression (Shirayama et al., 2011), but these
experiments have not definitively ruled out a role for the NAcc in antidepressant
functionality of endogenously-synthesised neurosteroids. We have therefore
examined inhibitory neurotransmission within the NAcc, and its response to
neurosteroid. The roles of the GABAA receptor α2 subunit in this response were
also assessed.
Chapter 1: Introduction 46
1.5. Thesis Aims
1.5.1. Generation of an α2Q241M knock-in mouse
Hydrophobic substitution of a critical α1Q241 residue (or equivalent in other α
subtypes) selectively disrupts neurosteroid binding to the potentiation site,
eliminating neurosteroid potentiation, with minimal effects on GABA activation
and receptor modulation by other allosteric ligands (barbiturates,
benzodiazepines) (Hosie et al., 2006; Hosie et al., 2009). We have therefore
used such a mutation to generate a transgenic knock-in mouse line that
harbours the mutation α2Q241M. By ablating neurosteroid potentiation at α2-type
GABAA receptors, this knock-in mutation allows us to examine the α2-subunit
specific functions of neurosteroids.
Especially because this mutation is influencing the response to an endogenous
compound, one must verify a lack of compensatory alterations in the knock-ins.
We have addressed this question using Western blot and quantitative
immunofluorescence assays to study GABAA receptor α subunit expression.
Because this mutation is not expected to alter diazepam or pentobarbital
sensitivity of mutant channels compared to wild-types, we can utilise these
compounds as positive controls within this study, to demonstrate that the effects
of GABAA receptor potentiation are intact within the transgenic animals.
1.5.2. Electrophysiological characterisation of α2Q241M mice
The roles of the α2 receptor isoform in the cellular responses to neurosteroids
have been examined with whole cell patch-clamp recordings in acute brain slice
tissue. The α2 subunit has been classically associated with synaptic events,
IPSCs (Prenosil et al., 2006), but we cannot rule out a role in extrasynaptic tonic
currents, and so both types of inhibition have been assessed. The areas chosen
Chapter 1: Introduction 47
for study, the hippocampus and NAcc, strongly express the α2 isoform, and are
key candidates for roles in anxiety and depression. To screen for effects of the
mutation on baseline neurotransmission, periods of control recording were
compared between wild-type and α2Q241M knock-in mice. Slices were also
treated with diazepam or THDOC, to evaluate the effect of the knock-in
mutation on sensitivity to these compounds.
1.5.3. Behavioural characterisation of α2Q241M mice
In this study, analyses have focussed on anxiety and depression-related
phenotypes, to assess the contribution of neurosteroids and α2-type GABAA
receptors to these disorders. We propose that neurosteroids may exert their
anxiolytic and antidepressant actions through α2 and α3-type GABAA receptors.
If our hypotheses are correct, homozygous α2Q241M mutant mice would be
predicted to show impaired anxiolysis and antidepression in response to
injected neurosteroid. Behavioural phenotypes of untreated mice would also
provide insight into the functions of endogenous neurosteroids acting at α2-type
GABAA receptors.
1.5.4. Summary of thesis aims
1. To establish that the α2Q241M mutation has the desired properties for this
study: loss of neurosteroid potentiation without altered GABA sensitivity
or benzodiazepine-mediated potentiation (Chapter 3).
2. To establish that α2Q241M knock-in mice have no compensatory changes
in receptor expression (Chapter 3).
Chapter 1: Introduction 48
3. To examine the consequences of α2Q241M knock-in for inhibitory
neurotransmission and its response to modulators (Chapter 4).
4. To define the GABAA-receptor α2-specific roles of endogenous and
injected neurosteroids in anxiety and depression-type behaviours
(Chapter 5).
Chapter 2: Materials and Methods 49
Chapter 2: Materials and Methods
2.1. Materials
2.1.1. Reagents
EDTA-Free Halt Protease Inhibitor Cocktail (100x) was purchased from Thermo
Fisher Scientific Inc. (Rockford, Illinois, USA); as were the bicinchoninic acid
(BCA) Assay Kit and the SuperSignal West Pico Chemiluminescent Substrate.
Protogel 30% w/v Acrylamide:0.8% w/v Bis-Acrylamide was purchased from
National Diagnostics (Atlanta, Georgia, USA). Protein Molecular Weight
Standards (Broad Range) were purchased from Bio-Rad Laboratories Ltd.
(Hemel Hempstead, UK). Polymerase Chain Reaction (PCR) was performed
using the Phusion Hot Start High-Fidelity DNA Polymerase kit (Thermo Fischer
Scientific Inc.) and primers from Eurofins MWG Operon (Ebersberg, Germany).
All other chemicals were obtained from Sigma-Aldrich (Steinheim, Germany) or
VWR International (Leuven, Belgium) unless indicated otherwise.
2.1.2. Antibodies
Table 2.1 provides details regarding sources and working dilutions for all
primary and secondary antibodies used in this study. The specificity of anti-
GABAA-receptor antibodies used in Western blotting was confirmed using
recombinantly expressed receptors. Briefly, Human Embryonic Kidney 293
(HEK293) cells were electroporated as described in Section 2.5.2. Two days
later, cells were washed in ice-cold phosphate-buffered saline (PBS; 137 mM
NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4) and scraped off
the culture dish in ice-cold RIPA buffer (150 mM NaCl, 20 mM Tris pH 8.0, 5
mM ethylene-diamine-tetra-acetic acid (EDTA), 1% v/v NP40, 0.5% w/v sodium
Chapter 2: Materials and Methods 50
deoxycholate, 0.1% w/v sodium dodecyl-sulphate (SDS), 1 mM phenylmethane-
sulfonylfluoride and 1x EDTA-Free Halt Protease Inhibitor Cocktail). Cells were
disrupted by rotating in the RIPA buffer (1 h at 4 oC); debris was removed by
sedimentation (30 min, 13,000 rotations per minute (r.p.m.), 4 oC, Eppendorf
5415R centrifuge). Further processing (BCA assay and SDS-polyacrylamide gel
electrophoresis (SDS-PAGE)) was performed as described in Section 2.3.2.
Each antibody tested produced a band between markers 45 and 66 kiloDaltons
(kDa), in the appropriate lane only (Fig. 2.2). Using the defined blotting
conditions (Section 2.3.2) and antibody dilutions (Table 2.1), it was therefore
evident that that these antibodies are GABAA α-subunit isoform specific.
2.2. Animals
All procedures involving animals were performed according to the Animals
(Scientific Procedures) Act, 1986 (ASPA) and had obtained local ethical
approval. When obtaining tissue for electrophysiology or Western blotting,
animals were decapitated under terminal isoflurane anaesthesia. After
behavioural testing, animals were culled by cervical dislocation according to
Schedule 1 of the ASPA. Given that the steroid hormones of the oestrus cycle
have a strong influence on brain neurosteroid levels in female mice (Corpechot
et al., 1997), only male mice were used in our experiments.
Chapter 2: Materials and Methods 51
Antibody
WB
dilution
IF
dilution Epitope Details Source
Prim
ary
An
tibo
die
s
mouse monoclonal anti-GABAA α1 1:1,000 - Amino acids 355-394 of mouse GABAA α1 NeuroMab 1
rabbit polyclonal anti-GABAA α1 - 1:20,000 Amino acids 1-16 of rat GABAA α1 gift from Dr Jean-Marc Fritschy 2
guinea-pig polyclonal anti-GABAA α2 - 1:1,000 Amino acids 1-9 of rat GABAA α2 gift from Dr Jean-Marc Fritschy 2
guinea-pig polyclonal anti-GABAA α2 1:1,000 - Amino acids 29-37 of rat GABAA α2 Synaptic Systems 3
rabbit polyclonal anti-GABAA α2 1:500 - Amino acids 1-10 of rat GABAA α2 gift from Dr Werner Sieghart 5
rabbit polyclonal anti-GABAA α3 1:1,000 1:1,000 Amino acids 29-37 of human GABAA α3 Alomone Labs 4
rabbit polyclonal anti-GABAA α4 1:4,000 1:500 Amino acids 1-14 of rat GABAA α4 gift from Dr Werner Sieghart 5
rabbit polyclonal anti-GABAA α5 - 1:500 Amino acids 1-12 of rat GABAA α5 gift from Dr Werner Sieghart 5
mouse monoclonal anti-β-Tubulin 1:1,000 - Tubulin from rat brain SigmaAldrich 6
(clone TUB2.1)
Se
con
da
ry
An
tibo
die
s
Alexa Fluor® 555 goat anti-guinea pig IgG (H+L) - 1:2,000 - Invitrogen Ltd. 7
Alexa Fluor® 555 goat anti-rabbit IgG (H+L) - 1:2,000 - Invitrogen Ltd. 7
HRP-conjugated goat anti-rabbit IgG (H&L) 1:10,000 - - Rockland Immunochemicals for Research 8
HRP-conjugated goat anti-mouse IgG(H&L) 1:10,000 - - Rockland Immunochemicals for Research 8
HRP-conjugated donkey anti-guinea-pig IgG(H&L) 1:1,000 - - Jackson ImmunoResearch Laboratories 9
Table 2.1 – Details of antisera employed in this project
Details of antisera utilised in this project for Western blot (WB) or immunofluorescence (IF). Secondary antibodies for WB were conjugated to Horseradish peroxidise (HRP).1
UC Davis, University of California, USA; 2 University of Zurich, Switzerland;
3 Goettingen, Germany;
4 Jerusalem, Israel;
5 Medical University Vienna, Austria;
6
Steinheim, Germany; 7 Paisley, UK;
8 Philadelphia, Pennsylvania, USA;
9 Stratech Scientific, Suffolk, England
Chapter 2: Materials and Methods 52
2.2.1. Generating GABAA receptor α2Q241M mutant mice
Mike Lumb, a molecular biology technician in the lab, performed this stage of
the project. Escherichia coli-based recombineering technology (obtained from
National Cancer Institute at Frederick, Maryland) was employed to create a
targeting vector containing exon 8 of GABAA receptor α2 subunit gene with
base pair changes for the point mutation Q241M (Fig. 2.1 Ai), utilising an RP23
Bacterial Artificial Chromosome library clone derived from a C57BL/6 mouse.
Using commercial facilities for homologous recombination (GenOway, Lyon,
France), the vector was introduced into 129Sv/Pas embryonic stem (ES) cells
by electroporation. Positive and negative selection procedures were used to
enrich for cells that had successfully undergone homologous recombination.
PCR and Southern blotting techniques were used to identify ES cell lines that
had precisely replaced the wild-type α2 sequence with that encoding the
Q241M mutant: the mutagenesis silently created an Nco-I restriction site,
allowing identification of homologous recombinants by Nco-I digestion followed
by Southern blot (band size decreases from 12 Kilobase-pairs (kb) to 6 kb (Fig.
2.1 B)). These ES cell lines were used to generate transgenic mice with the
α2Q241M mutation (Fig. 2.1 Aii). Germline-transmitted pups from a chimera x
C57BL/6J cross (F0 generation) were bred with C57BL/6J Cre-recombinase
expressing mice, in order to remove the neomycin resistance cassette (F1
generation; Fig. 2.1 Aiii). Further backcrosses were performed between
heterozygous (het) α2Q241M mice and C57BL/6J mice. The mice used in the
experiments described here are from generations F4-F6.
Chapter 2: Materials and Methods 53
Figure 2.1 – Generation of the mutant mice
A. i. Targeting construct comprising mouse genomic sequence (dark blue) containing
point-mutated exon 8 of GABAA α2 gene (red), together with positive and
negative selection markers Neo (neomycin resistance cassette, light blue) and
TK (Herpes simplex virus thymidine kinase, light blue). The Neo cassette is
flanked by loxP sites (green). This construct is housed in a low copy number
pBluescript KS+ plasmid backbone (pink).
ii. Result of successful homologous recombination; mutated exon 8 and neo have
been incorporated into genomic DNA (dashed dark blue line). TK and pBluescript
are lost. This corresponds to the genomic DNA from F0 generation mice.
Chapter 2: Materials and Methods 54
iii. Result of breeding recombinant mice with Cre-recombinase expressing mice;
Neo has been removed, leaving behind a single lox P site. Black arrowheads
mark locations of primer sites used for genotyping in C. This corresponds to
genomic DNA in mice from generation F1 and beyond.
B. Southern blot screen of ES cell lines: the Q241M mutation silently introduces a
novel restriction site – allowing a screen for restriction fragment length polymorphism.
Lanes 1-5, heterozygous recombinants; lanes 6-7, wild-type controls.
C. Primers and protocol for PCR genotyping of mouse genomic DNA, with
representative results achieved by running PCR products on a 2% w/v agarose gel.
D. Verification of the point mutation by DNA sequencing. Exon 8 was PCR-amplified
from mouse genomic DNA of wt (left), het (middle) and hom (right) animals, then
sequenced. Sequence results local to the point mutation are shown.
2.2.2. Breeding
Animals were maintained as het x het breeding pairs, to allow comparison of
homozygous (hom) knock-ins with wild-type (wt) littermates. Experimental
animals were males housed in cages of up to five littermates, with access to
food and water ad libitum, under a 12 h light-dark regime (lights on at 07:30h).
Males were housed with at least one other littermate post-weaning. For
electrophysiology and Western blotting, animals were used between postnatal
days 18 (P18) and 30. For behavioural characterisation and
immunofluorescence, animals aged between P42 and P70 were used.
2.2.3. Genotyping
Genomic DNA was isolated from ear clips and/or post-mortem tail snips by
incubating overnight in a lysis buffer (100 mM Tris-HCl pH 8.5, 5 mM EDTA,
0.2% w/v SDS, 200 mM NaCl and 0.1 µg/ml proteinase K (Roche Diagnostics
GmBH, Mannheim, Germany)) at 37 oC. Debris was removed by centrifugation
Chapter 2: Materials and Methods 55
(15 min, 13,000 r.p.m., room temperature (RT)); DNA was precipitated with
isopropanol (sedimented at 13,000 r.p.m., 5 min, RT), washed with 70% v/v ice-
cold ethanol (sedimented at 13,000 r.p.m., 5 min, RT), and re-suspended in
0.05x TE (10 mM Tris, 1 mM EDTA pH 8.0).
The mutant allele can be distinguished from the wt allele by the presence of the
loxP site remaining after removal of the neomycin cassette on the transgenic
allele (Fig. 2.1 Aiii). When PCR is performed as outlined in Fig. 2.1 C, the
presence of the lox P site results in a band 100 base pairs larger in the mutant
allele than the wt allele. A heterozygous mouse therefore produces two bands
(Fig. 2.1 C).
2.3. Western blotting
2.3.1. Protein isolation
Total protein was isolated from four brain areas from each animal: cortex,
cerebellum, hippocampus and nucleus accumbens (NAcc). The NAcc was
dissected from 350 µm coronal slices obtained as for electrophysiology (for
details, see Section 2.6.1). Remaining brain areas were isolated by direct
dissection from whole brain under a dissecting microscope, and the meninges
removed. Tissue was disrupted by homogenisation in ice-cold RIPA buffer,
either in a Dounce homogeniser (VWR International) (cortex, hippocampus,
cerebellum) or with a Microlance 25-gauge 1/2” needle and 1 ml syringe (BD
Franklin Lakes, NJ, USA) (NAcc). Cells were disrupted by repeated freeze-thaw
cycles, with debris removed after each cycle by centrifugation (20 min, 13000
r.p.m., 4 oC).
Chapter 2: Materials and Methods 56
2.3.2. Polyacrylamide gel electrophoresis (PAGE) and blotting
Protein concentration was determined spectrophotometrically (Bio-Rad
SmartSpec Plus) according to instructions with the BCA assay kit. Proteins
were denatured at RT by addition of Laemmli Sample Buffer (150 mM Tris-Cl
pH 6.8, 6% w/v SDS, 0.3% w/v Bromophenol blue, 30% w/v glycerol and 15%
v/v β-mercaptoethanol) and run on a 10% SDS-PAGE gel (5% stacking gel) in
Running Buffer (25 mM Tris, 192 mM glycine, 0.1% SDS) for 1 h at 150 V, using
a Bio-Rad Miniprotean II system. Proteins were transferred to nitrocellulose
membranes (Hybond C Extra, Amersham Biosciences, Buckinghamshire, UK)
at 25 V for 70 min in Transfer Buffer (25 mM Tris, 192 mM glycine, 20% v/v
methanol) using an XCell II Blot Module (Invitrogen, Carlsbad, USA).
Successful transfer was confirmed by Ponceau staining (0.1% w/v Ponceau-S
in 5% v/v acetic acid); the stain was thoroughly washed off the membrane with
PBS before the blocking and incubation steps.
All blocking, washing and antibody incubation steps were performed on a
shaker with 4% milk in PBS (plus 0.1% TWEEN-20), the antibody dilutions
employed are outlined in Table 2.1. Blocking was performed for 1 h at RT,
whilst primary antibody incubation was overnight at 4 oC, and secondary
antibody incubation was for 2 h at RT. After each antibody incubation,
membranes were washed three times (20 min, RT). Blots were finally rinsed in
PBS and developed using chemiluminescent substrate and visualised with an
ImageQuant LAS4000 imager. Images were quantified using the Western blot
plug-in on ImageJ software (Version 1.44p, National Institutes of Health, USA).
After blotting for the relevant GABAA α subunit, blots were subjected to a mild
stripping procedure (10-20 min incubation in a buffer comprising 200 mM
glycine, 0.1% w/v SDS, 1% v/v TWEEN-20, pH 2.2) followed by 6 x 5min
washes in PBS. Successful stripping was confirmed by re-incubation in
secondary antibody (1 h at RT), washing (3 x 20 min, RT) and developing,
where no residual signal was observed. Membranes were then re-blotted for
quantification of β-tubulin expression. The density of each α subunit band was
Chapter 2: Materials and Methods 57
first normalised to its corresponding β-tubulin band, then signals within each
Western blot were normalised to the average signal of the wt bands in that blot
(such a normalization procedure has been used by others – e.g. Chandra et al.,
2005).
Before quantitative blots were carried out, titration experiments were performed
to determine the amount of total protein, which, when blotted under the above
conditions, produces a signal in the middle of the dynamic range of the
detection system. When working under such conditions, one maximises the
chances of detecting any up- or down-regulation of GABAA subunit expression.
Gels were therefore loaded with increasing amounts of total protein (ranging
from 5 to 100 µg) from each brain area from three P21-P30 C57BL/6J mice and
blotted for each GABAA α subunit. These titration curves (Fig. 2.2 and data not
shown) were used to select an appropriate amount of protein to load when
quantifying subunit expression from the transgenics (Table 2.2).
Subunit Cortex Hippocampus
Nucleus
Accumbens Cerebellum
α1 15 µg 25 µg 25 µg 20 µg
α2 25 µg (gp) 25 µg (gp) 20 µg (r) n.d.
α3 45 µg 25 µg 20 µg n.d.
α4 30 µg n.d. n.d. n.d.
Table 2.2 – Amount of total protein loaded for Western blotting
Results from titration experiments determining the amount of total protein, which,
when blotted under the defined conditions, produces a signal in the middle of the
dynamic range of the detection system. Two different antisera were used in
detecting α2 expression – guinea pig (gp) or rabbit (r) as indicated. n.d. - not
determined because no reliable signal was obtained with calibration blots (0-100
µg total protein loaded).
Chapter 2: Materials and Methods 58
Figure 2.2 – Optimising conditions for Western Blotting
A. Antibodies are specific for their defined α subunit: 20 µg of each HEK293 cell lysate
was prepared and blotted as described in Section 2.1.2. Images show the blot between
66 kilodalton (kDa) and 45 kDa markers. Antibodies: α1, mouse monoclonal anti-α1;
α2(gp), guinea-pig polyclonal anti-α2; α2(r), rabbit polyclonal anti-α2, α3, rabbit
polyclonal anti-α3; α4, rabbit polyclonal anti-α4.
B. Calibration curve for the signal (arbitrary units) obtained by blotting increasing
amounts of C57BL/6J cortical protein for α1; average of 3 blots (error bars = standard
error of the mean (s.e.m.)). Loading 15 µg total protein from test cortices will give a
mid-range signal. Inset: representative example blot (bands 5-50 µg correspond to the
points on the curve below). Similar calibration curves were obtained for other subunits
and other brain areas.
2.4. Immunofluorescence
2.4.1 Brain sectioning for immunofluorescence
Three mice of each genotype (wt, het, hom) were sacrificed with a lethal dose of
sodium pentobarbital (dissolved in 0.9% w/v NaCl to a concentration of 15
mg/ml, injected at a volume of 10 ml per kg body weight), and subjected to
cardiac perfusion first with 20 ml ice-cold saline/heparin mix (5000 units heparin
Chapter 2: Materials and Methods 59
per litre (Leo Laboratories Ltd., Buckinghamshire, UK), 0.9% w/v NaCl), then
with 10 ml ice-cold fixative (4% paraformaldehyde in 0.1 M Phosphate Buffer
(PB: 12 mM NaH2PO4, 38 mM Na2HPO4)). Brains were removed, and incubated
for a further 2 h at 4 oC in the fixative. Finally, they were transferred to a
cryoprotectant solution comprising 0.1 M PB, 0.03% w/v sodium azide and 30%
w/v sucrose for incubation at 4oC until tissue sank (at least overnight).
Coronal 40 µm slices were cut from frozen brain on a Leica SM200R sliding
microtome (Leica Microsytems GmBH, Wetzlar, Germany). Sections were
collected in a 24-well plate (Starstedt Ltd., Leicester, UK), filled with 0.1M PB
with 0.03% w/v sodium azide, and were stored at 4 oC until staining. During
sectioning, serial sections were inserted into adjacent wells such that one well
contains every 6th slice of the nucleus accumbens, or every 12th slice of the
hippocampus. When staining, sections were taken from the same well, thus
ensuring a uniform representation along the anterior-posterior axis of the
structure being examined.
2.4.2. Staining sectioned tissue
Tissue sections were rinsed twice with PBS before incubating for 1 h at RT in
permeabilisation/blocking solution (0.5% bovine serum albumin, 2% normal
goat serum, 0.2% triton-x-100 in PBS). Sections were incubated overnight at 4
oC with primary antibody dissolved in the same permeabilisation/blocking
solution. Secondary antibody was applied to slices at RT for 2 h. After each
antibody incubation, slices were washed four times in PBS (15 min at RT). All
incubations and washes were performed by gently shaking slices in 24 well-
plates. Stained slices were mounted on super premium glass microscope slides
(VWR International) using ProLong Gold Antifade reagent (Invitrogen Ltd.
(Paisley, UK)). Slides were stored in a dark container at 4 oC, until imaging.
Controls for non-specific labelling by secondary antibodies were performed by
staining as described, but omitting primary antibody.
Chapter 2: Materials and Methods 60
2.4.3. Image acquisition and analysis
Image acquisition was performed using a Zeiss Axioscop LSM510 confocal
microscope (Carl Zeiss Ltd., Welwyn Garden City, Hertfordshire, UK), equipped
with three laser lines (λ = 488, 543 and 643 nm) and Plan Neofluor 20x air
(numerical aperture (NA) 0.5) and 63x oil (NA 1.4) differential interference
contrast objectives (Carl Zeiss). Images were captured as z-stacks, with images
in each plane acquired as a mean of 8 scans in 8 bits and stored for analysis.
For each subunit and brain area, three slices were imaged per animal. Image
acquisition and analysis were performed blind to the genotype of the animal.
Images of the cornu ammonis 1 (CA1) region of the hippocampus were
acquired with the 63x lens at 1x zoom, and encompass the cell body layer and
the apical dendrites. Images of the dentate gyrus (DG) were also acquired with
the 63x lens at 1x zoom, and encompass the granule cell layer and molecular
layer of the medial blade of the DG. With the exception of images for α3
expression, images for the NAcc were acquired with the 63x lens at 1x zoom,
and represent both core and shell regions. Because the expression of α3 was
much less uniform throughout the NAcc, showing patches of intense staining in
both core and shell regions, images for this subunit were instead acquired using
the 20x lens at 1x zoom. This reduces the spatial resolution of the images, but
acquires data over a 9.9 times larger surface area, and so compensates for this
‘patchy’ distribution.
Images were analysed using ImageJ (Version 1.44p). For CA1 and DG, the
mean fluorescence intensity was determined for separate regions of interest
(ROIs) encompassing the dendrites and cell bodies. For the NAcc, mean
fluorescence intensity was measured for the entire image, rather than a specific
ROI. In both cases, to ensure that the value is representative of the staining
throughout the brain slice, three readings were taken per image: one from the
Chapter 2: Materials and Methods 61
top, one central and one at the bottom of the z-stack; the average of these
values was taken to represent the expression of that subunit in that slice.
Bar charts in Fig. 3.5 represent the average mean fluorescence intensity across
the whole collection of slices. Results are expressed relative to values for wt
animals, such that het or hom values of 1.0 would represent no change in
expression; and, for example 1.2 would represent a 20% increase, or 0.8 a 20%
decrease in expression.
The mean fluorescence intensity approach is less appropriate in images where
punctae are in the minority against background (e.g. see α1 expression in the
NAcc, Fig. 3.4 C). In this case, and also for α3 expression in the DG (where the
mean fluorescence intensity analysis indicated a tendency toward a difference),
subunit expression was examined in more detail by quantifying immunopositive
punctae. This was carried out using the FociPicker3D plug-in in ImageJ
(Guanghua Du, Institute of Modern Physics, CAS, China), which searches for
local maxima in three-dimensions. As demonstrated in Fig. 2.3 A, when using
optimal settings (intensity and size thresholds to define positive foci against the
local background), the program successfully picks out immunopositive punctae,
and provides measurements of their volume and intensity.
The FociPicker3D approach was not extended to other images for several
reasons. The plug-in is not suitable for images containing a mixture of diffuse
and punctate staining (e.g. α4 and α5 subunit immunofluorescence), because it
fails to distinguish between a small punctae and diffuse immunofluorescence
(see Fig. 2.3 B). The bright but diffuse staining of some dendritic processes in
hippocampi stained with α1-selective antiserum poses a similar problem (being
recognised as a large number of small punctae by FociPicker3D). In addition,
no settings were found that enabled the FociPicker3D plug-in to satisfactorily
identify α2 immunopositive punctae in any of the regions imaged, probably
because of the high background fluorescence in these images (which may
represent diffuse staining for this subunit).
Chapter 2: Materials and Methods 62
Figure 2.3 – FociPicker3D successfully identifies immunopositive punctae if
images lack diffuse staining
A. Series of images demonstrating good correlation between immunopositive punctae
for α3 staining in the dentate gyrus (top) and punctae identified by FociPicker3D
(bottom). Images on the left represent the entire field of view (63x zoom), middle
images represent a zoom into the yellow box, images on the right are zoomed into the
red box from middle images. With appropriate threshold settings, FociPicker3D
identifies the majority of immunopositive punctae in the image, and rarely picks up
background fluorescence as false positives. Scale bars (white) are 30 µm.
B. Representative example for α5 immunostaining in the CA1 region of the
hippocampus (raw image on left, Foci Picker result on the right). The program struggles
to identify true punctae in images that contain diffuse as well as punctate staining.
Some regions of diffuse staining wrongly identified as regions of many small individual
punctae.
Chapter 2: Materials and Methods 63
2.5. HEK293 cell culture and electrophysiology
2.5.1. HEK293 cell culture
HEK293 cells were maintained on 10 cm plates (Greiner-Bio-One GmbH,
Frickenhausen, Germany) in a culture medium comprising Dulbecco’s modified
Eagle medium supplemented with 10% v/v fetal bovine serum, 100 µg/ml
streptomycin, 100 U/ml penicillin G and 2 mM glutamine (all from Gibco,
Invitrogen Ltd.). Plates were incubated at 37 oC in 95% air/5% CO2 (BOC
Healthcare, Manchester, UK). Cultures were passaged at approximately 70-
80% confluency and plated at appropriate dilution either on to 10 cm plates for
maintenance, poly-L-lysine (100 µg/mg)-coated 18 mm coverslips (VWR
international) for electrophysiology, or 6 cm plates (Nunclon-Δ Surface, Thermo
Fisher Scientific Inc.) for biochemistry.
For passage, cells were washed with 5 ml Hank’s balanced salt solution
(HBSS), and detached by trypsinisation with 2 ml 0.05% w/v trypsin-EDTA
(Gibco). Cells were resuspended in 10 ml culture medium, which quenches the
trypsin, and then pelleted by centrifugation at 1000 r.p.m. for 5 min. Supernatant
was removed, and the cell pellet re-suspended in 5 ml culture medium by
trituration with a fire-polished glass Pasteur pipette (VWR International), to
ensure a single-cell suspension before plating.
2.5.2. HEK293 cell transfection
For electrophysiology, HEK293 cells were transfected by a calcium phosphate
precipitation method. After plating on coverslips, cells were exposed to a
mixture comprising 1 µg of each cDNA expression vector encoding the required
GABAA receptor subunits (murine α, β, γ or δ cDNA housed in a pRK5 plasmid
Chapter 2: Materials and Methods 64
vector), 1 µg of cDNA expression vector encoding enhanced green-fluorescent
protein (eGFP), 20 µl 340 mM CaCl2 and 24 µl 2 x HBSS (280 mM NaCl, 2.8
mM Na2HPO4, 50 mM HEPES, pH 7.2). Cells were subjected to
electrophysiology 18-48 h later.
For Western blotting, cells were prepared as described in Section 2.5.1., except
that after trypsinisation, cells were pelleted and washed by resuspending in 10
ml OptiMEM I (Gibco) and again sedimented by centrifugation. This cell pellet
was resuspended in 400 µl OptiMEM I per transfection and transferred to
BioRad 0.4 cm electrode gap cuvettes containing the appropriate cDNA vector
mixture (2 µg per GABAA receptor subunit). Cells were then electroporated
using a BioRad gene pulser II electroporator, and recovered with 0.5 ml culture
medium before plating on poly-L-lysine-coated 6 cm culture dishes. Protein was
harvested from cells 2-3 days post-electroporation.
2.5.3. HEK293 cell electrophysiology
Coverslips containing transfected HEK293 cells were transferred to a recording
chamber, and continuously perfused at RT with Krebs solution, containing
(mM): 140 NaCl, 4.7 KCl, 1.2 MgCl2, 2.5 glucose, 11 HEPES and 5 CaCl2 (pH
7.4). Whole-cell recordings were undertaken from transfected (GFP-positive)
HEK293 cells located using epifluorescence optics (Nikon Eclipse E600FN,
Nikon Instruments Europe B.V. Surrey, UK). Membrane currents were recorded
using a Multiclamp 700A amplifier (Molecular Devices, Sunnyvale, California,
USA) in the voltage-clamp configuration (holding potential of -20 mV). Currents
were filtered at 4 kHz (8th order Bessel, -48 dB/octave) and digitized at 50 kHz
(Digidata 1322A, Molecular Devices), and displayed using Clampex software
(version 8.2, Molecular Devices). Patch pipettes (4-5 M ) were filled with an
internal solution containing (mM): 120 KCl, 1 MgCl2, 11 EGTA, 10 HEPES, 1
CaCl2, 4 ATP (pH 7.2). The osmolarity of the internal solution was 300 ± 20
miliOsmoles/litre (mOsm/l), measured using a vapour pressure osmometer
Chapter 2: Materials and Methods 65
(Wescor Inc, Utah, USA). Responses of cells to brief (2-4 s) applications of
GABA, alone or in combination with other drugs, were recorded using the gap-
free recording mode in Clampex. Substance applications were made using a Y-
tube (Fig. 2.4), and with 2 min recoveries between applications.
Figure 2.4 – Schematic representation of
the Y-tube
Auxiliary/wash and Y-tube/drug-application
tubes are arranged as depicted. Between
drug applications, solenoids are open –
allowing Krebs to flow over the cell, and drug
to flow to waste under vacuum pressure.
During applications, solenoids close, allowing
drug to be applied to the cell (dotted line) in
the absence of washing Krebs.
Responses were measured by comparing peak amplitude to the baseline
holding current in Clampfit software (version 10.2, Molecular Devices). To
monitor for run-up or run-down of GABA responses, a defined concentration of
GABA was applied at regular intervals to the same cell. Between applications,
series resistance (Rs) was monitored. Experiments were terminated if Rs
changed by more than 20%. GABA concentration-response curves were
constructed by plotting mean peak responses against the GABA concentration,
and fitting these data using the Hill equation (equation 1).
Equation 1
I[GABA] = Imax * ([GABA]n / (EC50n + [GABA]n))
Where I[GABA] is the peak current activated by GABA at a concentration [GABA],
Imax is the maximal current response to GABA, EC50 is the concentration of
GABA producing a current response 50% of the maximal response, and n is the
Hill coefficient.
Chapter 2: Materials and Methods 66
GABA concentration-response curves (Fig. 3.1) demonstrate that 0.5 µM GABA
corresponds to an EC15 concentration (producing 15% of the maximal
response) for both α2WTβ3γ2s and α2Q241Mβ3γ2s combinations. This
concentration was selected for determining potentiation by THDOC and/or
diazepam. The peak response to 0.5 µM GABA + drug was divided by that
produced by a preceding control application of 0.5 µM GABA alone. The
diazepam potentiation curve was fit with a modified version of equation 1, to
include the pedestal (equation 2).
Equation 2
P[dzp] = Pmin + (Pmax – Pmin) * ([dzp]n / (EC50n + [dzp]n))
Where P[dzp] is the peak potentiation activated by 0.5 µM GABA plus a particular
concentration of diazepam ([dzp]), Pmax is the maximal potentiation achieved
with diazepam, EC50 is the concentration of diazepam producing 50% of the
maximal potentiation, and n is the Hill coefficient.
2.6. Brain slice electrophysiology
2.6.1. Preparation of slices
Coronal slices (350 µm) were obtained from wild-type and transgenic mice
using a Leica VT1200s vibroslicer (Leica Microsytems GmBH, Wetzlar,
Germany). Sections of nucleus accumbens were cut in an ice-cold artificial
cerebrospinal fluid (aCSF) comprising (mM): 85 NaCl, 2.5 KCl, 1.25 Na2H2PO4,
26 NaHCO3, 1 CaCl2, 4 MgCl2, 25 glucose, 75 sucrose and 2 kynurenic acid.
Hippocampal slices were cut in ice-cold aCSF comprising (mM): 130 K-
gluconate, 15 KCl, 0.05 EGTA, 20 HEPES, 4 Na-pyruvate, 25 glucose and 2
kynurenic acid, pH 7.4). Slices were transferred to a holding chamber in a water
bath set at 37 oC, whilst the solution was slowly exchanged over 1 h to a
Chapter 2: Materials and Methods 67
recording aCSF of composition (mM): 125 NaCl, 2.5 KCl, 1.25 Na2H2PO4, 26
NaHCO3, 2 CaCl2, 1 MgCl2, 25 glucose. Slices were then maintained in the
holding chamber at RT. All aCSF solutions were continuously bubbled with
95%O2/5%CO2 (BOC Healthcare).
2.6.2. Whole-cell patch clamp recording
Whole-cell recordings were undertaken at RT from single neurons located using
infra-red optics (Nikon Eclipse E600FN, Nikon Instruments Europe B.V. Surrey,
UK) fitted with a Basler SLA750-60fm Camera (Basler Vision Technologies,
Ahrensburg, Germany). Membrane currents were recorded as described for
HEK293 cells (Section 2.5.3), using patch pipettes (3.8-4.5 M ) filled with an
internal solution containing (mM): 140 CsCl, 2 NaCl, 10 HEPES, 5 EGTA, 2
MgCl2, 0.5 CaCl2, 2 Na-ATP, and 2 QX-314 Bromide (Ascent Biochemicals,
Abcam plc, Cambridge, UK). Solution osmolarity was adjusted with sucrose,
usually 5 mM for hippocampal slices (approximately 305 mOsm/l) and 12.5 mM
for NAcc slices (approx. 315 mOsm/l).
Slices were continuously perfused with recording aCSF supplemented with
kynurenic acid, to isolate GABAergic events. The neurosteroid, THDOC was
dissolved as a 10 mM stock solution in dimethyl sulphoxide (DMSO), and
diazepam was dissolved as a 100 mM stock in DMSO. These drugs were
diluted to the appropriate final concentration in the recording aCSF, and applied
to cells in the bath after a period of stable control recording. Drugs were allowed
to equilibrate in the bath for at least 5 min; control or ‘mock’ recordings were
performed to confirm that there are no changes in synaptic currents due to the
DMSO vehicle, or washout of intracellular contents independent of drug (i.e.
cells were challenged with 0.01% (v/v) DMSO1, or simply held in control aCSF
1 Equivalent to the highest % DMSO applied to cells (when administering 100 nM THDOC)
Chapter 2: Materials and Methods 68
for an equivalent duration to experiments where drug was applied)2.
Experiments were completed by the bath application of 20 µM (-)-bicuculline-
methiodide: to confirm that all events were GABAergic, and to allow any GABA-
mediated tonic currents to be measured. Recordings were made in 2 min
epochs, between which Rs was monitored, and experiments were terminated if
Rs changed by more than 20%.
2.6.3. Analysis of spontaneous inhibitory post-synaptic currents (sIPSCs)
To determine the frequency of sIPSC events in a given recording, all visible
events in an epoch were identified using MiniAnalysis software (Synaptosoft
Inc., Fort Lee, New Jersey, USA), and inter-event interval times determined. Not
all events are appropriate for further analysis because, for example, their decay
is contaminated with the presence of another sIPSC event. Only clean,
uncontaminated events were selected for decay fitting in MiniAnalysis, using
either a mono- or bi-exponential function. The best fit was determined by eye
and also by the increasing value of the parameter R2. In order to combine
analysis of mono- and bi-exponentially decaying events, decay times were
transformed to a weighted decay time, w, according to equation 3.
Equation 3
w = (A1. 1 + A2. 2) / (A1 + A2);
Where 1 and 2 are the time constants for a biexponential decay, and A1 and A2
are the relative amplitude contributions of 1 and 2, respectively. For
monoexponential decay, terms A2 and 2 are absent (i.e. w = 1).
For each IPSC event, four parameters were measured: rise time, amplitude, w
and area (charge transfer). A minimum of 100 events per condition (control vs.
2 Both controls gave equivalent results so have been combined
Chapter 2: Materials and Methods 69
drug) were fitted, and average sIPSC parameters were determined. The results
are expressed as a percentage change for the test epoch relative to the control
epoch in that cell; i.e. +30% and -30% would represent a 30% increase and
decrease, respectively, for a particular parameter.
2.6.4. Analysis of tonic GABA currents
Tonic currents are revealed by observing a reduced holding current and root
mean square (r.m.s.) current noise on application of the competitive GABAA-
receptor antagonist, 20 µM bicuculline. Changes in holding current were often
small (less than 10 pA) and changes in r.m.s. noise proved to be a more robust
and reliable measure of changes in tonic current in this study. WinEDR (version
3.1) software (John Dempster, University of Strathclyde, Glasgow, UK) was
employed to measure r.m.s. noise over 100 ms epochs of a recording. The
presence of any synaptic events in a single 100 ms epoch will increase the
r.m.s. value, and so such epochs must be excluded from analysis. It is possible
to do this manually by scrolling through records and marking those
contaminated epochs for exclusion, but we used an automated procedure.
Briefly, Microsoft Excel was employed to compare the r.m.s. value of each 100
ms epoch to a user-defined threshold, and automatically exclude those with
values above the threshold as ‘contaminated’. The threshold is defined as a
proportion of the median r.m.s. current over a local 5 s window of recording (50
x 100 ms epochs). This ‘thresholding approach’ was validated by comparison
with a small section of data (approx. 60 s) where contaminated 100 ms epochs
were excluded manually. This threshold was then applied to the remainder of
the data from that cell. For both tonic and synaptic current analyses, time
courses indicated that the full effect of drug was achieved within 5 min, and so
data report the changes in r.m.s. noise or IPSC parameters after at least 5 min
equilibration of drug.
Chapter 2: Materials and Methods 70
2.7. Behavioural Analyses
2.7.1. Animal handling and drug administration
All experimental procedures were performed between 07:30h and 12:30h.
Animals were transferred from their housing facility to the experimental room 1
hour before testing commenced. During this habituation time, mice were housed
individually in cages, with access to ample food and water. All animals were
used in a single behavioural test, and were experimentally naïve. Before
behavioural testing, mice were subjected only to routine handling for husbandry
and ear notching.
Where drugs were administered, mice received substances by intraperitoneal
injection (27 gauge ½” needle, 1 ml syringe) at a volume of 10 ml per kg body
weight. The timings of injection were either 15 min (THDOC, sodium
pentobarbital) or 30 min (diazepam) before behavioural testing. THDOC was
dissolved in 0.9% w/v NaCl containing 25% w/v 2-hydroxypropyl-β-cyclodextrin
vehicle. Diazepam was prepared in 0.9% w/v NaCl containing 3.5 % v/v DMSO
vehicle. Sodium pentobarbital was dissolved directly in 0.9% w/v NaCl. Dose-
ranges and timings for injection were determined by examining those used
elsewhere in the literature and by performing preliminary experiments with
these mice (e.g. timecourses for the action of THDOC elsewhere (Crawley et
al., 1986) and here (Fig. 2.8) indicate that anxiolytic and motor-impairing effects
peak within 15 min post intraperitoneal injection).
2.7.2. Elevated plus maze
The elevated plus maze is a standard screen for anxiety-like behaviour,
validated with anxiolytic drugs (Pellow et al., 1985; Pellow & File, 1986). The
Chapter 2: Materials and Methods 71
maze comprised a horizontal black Perspex cross, with a central square of 50 x
50 mm, from which extend four arms, each to a length of 300 mm (Fig. 2.5).
Two ‘closed’ arms were surrounded by black Perspex walls (height 150 mm),
and the other two ‘open’ arms had only a raised edge (approx. 2 mm). The
maze was elevated 500 mm above the floor of the experimental room, and the
illumination across the arms was uniform and low intensity (approximately 25
lux, equivalent to 25 lumen per square metre).
Mice were placed in the centre of the maze, initially facing one of the closed
arms. Their movements within the maze over a 5 min trial were monitored with
a Sony Handycam (HDR-CX-190E) camcorder mounted 1.5 m directly above
the maze. Between tests, the maze was wiped down with distilled water, and
allowed to dry. The number of entries into each arm and the time spent in each
arm was scored blind to genotype. A mouse was defined as having entered an
arm when all four paws were within the boundary of the arm. The percentage
time in an arm is expressed as (time in arm / total time of experiment) x 100.
Figure 2.5 – The elevated plus maze
The maze is depicted from plan (top)
and side elevation views (bottom). The
closed arms (C) have black Perspex
walls 15 cm high, whilst open arms (O)
have only a slight raised edge. The
maze is elevated by mounting on a
tripod, and behaviour on the maze is
scored from a video captured by a
camera mounted above the maze.
Chapter 2: Materials and Methods 72
For examining baseline anxiety levels, three mice were tested each day – one
of each genotype – in a randomized order. When examining the anxiolytic
effects of drugs, mice were tested in pairs (one hom and one wt), where the
drug dose received, and the order in which mice were tested (wt first vs. hom
first) had been randomised. Videos were scored blind to both the genotype and
drug administered.
Note that whilst some investigators include a series of ethological measures of
‘risk assessment’ in their anxiety studies (Rodgers & Johnson, 1995, 1998), we
have not included this analysis here for a number of reasons. There is a degree
of subjectivity and ambiguity when scoring particular risk assessment
behaviours, not aided by our study using a black mouse, which is filmed on a
black background under dim lighting. Moreover, it is difficult to say what a
reduction in risk assessment behaviour actually represents: reduced anxiety,
increased anxiety or activity effects (Blanchard et al., 1990). We have therefore
focussed on spatiotemporal scores for anxiety (percentage time on open arms
etc.), which respond more consistently to a range of anxiolytics than ethological
measures (Rodgers & Johnson, 1998).
2.7.3. Light-Dark Box
The light-dark box is a standard screen for anxiety-like behaviour, with validity
for anxiolytic drugs (Crawley & Goodwin, 1980). The box dimensions were 20 x
45 cm, which were divided into a light zone (850 lx) and a dark zone (50 lx)
connected by a small doorway. The light zone has an area twice that of the dark
zone and the floors of each zone are marked with a grid to allow scoring of
activity (Fig. 2.6). A mouse was placed initially in the light zone, facing away
from the doorway, and was then allowed to freely explore the box for 10 min.
Three mice did not move for at least the first 5 min of the test, and have been
excluded from the analysis. Between tests, the box was wiped clean with
distilled water, and allowed to dry.
Chapter 2: Materials and Methods 73
Figure 2.6 – The light-dark box
Plan view is shown. The light zone
consists of white Perspex floor and walls
(25 cm high), and the dark zone of black
Perspex. Zones are connected by a small
doorway at floor level. The floor of each
zone is divided into equally-sized squares
(5 x 5 cm) to allow activity to be scored
(see main text).
To compare the effects of genotype on baseline anxiety, four mice were tested
each day (two wt and two hom) – in a randomised order. To compare the effect
of drug treatments, mice were tested in pairs – one wt and one hom; the order
in which they were tested (wt first, or hom first), and the drug dose received,
were randomised.
A video record of the test was scored blind to the genotype and any drug
treatment of the mouse. Scoring assessed the time to leave the light zone, the
time spent in each zone, the number of transitions between each zone. In
addition, the exploratory activity within each zone was measured by counting
the number of grid-lines crossed, then correcting for the amount of time spent in
each zone (i.e. activity is expressed as line crossings per second).
2.7.4. Tail-suspension test
The tail suspension test is a standard screen for depression, with validity for
anti-depressant drugs (Steru et al., 1985; Thierry et al., 1986). Mice were
Chapter 2: Materials and Methods 74
suspended by the tail with 19 mm wide PVC insulation tape (Powerlink Plus+,
Lancashire, UK) from a metal bar, raised 55 cm above the benchtop (see Fig.
2.7). Because of the tendency of the C57BL/6J strain to climb their tail in this
test, the ‘climbstopper’ approach described by Can et al. (2012) was used – i.e.
a small plastic cylinder (4 cm long, 1 cm internal diameter, approximately 1
gram weight) was placed over the base of the tail before attaching the sticky
tape. With the equipment depicted in Fig. 2.7, four mice can be tested
simultaneously. The mice cannot observe or interact with each other due to the
opaque dividers between compartments. Two wt and two hom animals were
tested together, with the drug dose applied, and compartment in which they
were placed, determined in a counterbalanced, randomised design. The
behaviour of mice was filmed over a 6 min period.
The scoring was performed blind to the genotype and drug dose applied, using
an Xnote stopwatch (dnSoft Research Group) to count the cumulative time for
which the mouse is immobile over the six minutes of the test (Can et al., 2012).
Small movements of front paws, without hind leg movements, are not counted
as escape-related mobility; similarly pendulum-like motion, resulting from
momentum gathered in previous bouts of motion, is not counted as mobility. An
increased immobility can represent a depressed phenotype, although care must
be taken to ensure this is not a sedative effect. Conversely a decrease in
immobility is thought to represent a reduction in depression, again taking care to
ensure this is not a stimulant effect.
Chapter 2: Materials and Methods 75
Figure 2.7 – Tail suspension
test equipment
A white box 60 cm in height is
divided into four compartments,
13 cm wide and 11 cm deep.
Mice are suspended from an
aluminium bar (grey), elevated
55 cm above the bench top,
using sticky tape (blue). Before
attaching to the bar, tails are
also fitted with a climbstopper
(see main text). Dividers
between compartments (black
lines) prevent mice from
observing one-another during
the test.
2.7.5. Rotarod
The rotarod test is a standard screen for motor coordination defects induced by
genetic mutation or drug treatment of rodents (Dunham & Miya, 1957; Jones &
Roberts, 1968). Mice were placed on a five-station rotarod treadmill apparatus
(Med Associates Inc., Vermont, USA) with the rod rotating at an initial speed of
4 r.p.m. The rod was then programmed to accelerate uniformly over a period of
5 min up to a final speed of 40 r.p.m. If a mouse failed the test by falling off the
rod before the full 5 min trial, it was removed from the apparatus until the next
trial. The time at which a mouse fell from the rotarod was recorded by the
MedAssociates software, triggered by a photocell beam break. In some
instances, a mouse will hold on to the rod and rotate with it – termed a ‘passive
rotation’ – this is also considered a failure; the time at which this occurred was
noted manually. If a mouse completed the full test, the ‘time to failure’ was
simply set as the duration of the test (300 s).
Chapter 2: Materials and Methods 76
The performance of each mouse on his first encounter on the rod (green bar,
Fig. 2.8) was used as a measure of the baseline motor coordination. Each
mouse was subjected to further training on the rotarod, such that a consistent
performance was attained before the effect of drug was assessed. Training
consisted of 5 trials on the rotarod per day, with a 10 min break between trials
(Fig. 2.8); training was carried out over four consecutive days. On the fifth day,
each mouse mice was injected intraperitoneally with THDOC (or vehicle) 15 min
before his first exposure to the rod (arrow, Fig. 2.8). The drug dose
administered to each mouse was pre-determined by a counterbalanced
randomisation procedure.
To assess any impairment of motor coordination caused by a drug, the
performance of each mouse under drugged conditions was expressed relative
to that mouse’s average performance on the rod on the preceding day (i.e.
average of blue trials, Fig. 2.8). Each mouse was tested five times under the
influence of drug, once every 15 min, starting from 15 min after injection. The
recovery from any motor impairment induced by THDOC was fast (Fig. 2.8B),
therefore the effect of THDOC was expressed as the time to failure on the first
trial after injection, normalised to the trials on day 4 of training.
Chapter 2: Materials and Methods 77
Figure 2.8 – Rotarod protocol
A. The rotarod protocol; each downward bar represents a 5 min trial on the
accelerating rotarod, mice were rested 10 min between exposures. The first exposure
(green) gives a measure of baseline motor coordination. On the testing day, mice were
injected (red arrow) 15 min before exposure to the rotarod. The performance under the
influence of drug (red) is normalised to the average performance under non-drugged
conditions after training (blue).
B. The motor-impairing effects of 20 mg/kg THDOC were short-lived, and most clearly
apparent in the first trial (15 min after injection); recovery to baseline was achieved by
trial 3 (45 min after injection). Comparisons therefore focus on the trial-1 effects of
THDOC.
2.8. Statistics
All data are presented as mean ± standard error of the mean (s.e.m.). Graphical
representations of data were plotted using Microcal OriginPro (version 8.5;
Chapter 2: Materials and Methods 78
OriginLab Corporation, Northampton, MA, USA), and figures assembled using
CorelDraw (version X4; Corel UK Ltd., Maidenhead, Berks, UK).
For most of the data analyses, parametric comparisons (analysis of variance
(ANOVA) and t-test) were first considered for use. These tests assume that the
data are normally distributed, and that the variances of the compared groups do
not differ. Tests were performed to verify that these constraints were met (e.g.
Bartlett’s test to compare standard deviations of data sets). Where the data did
not satisfy these conditions, they were variously transformed (e.g. logarithmic
transform), and the criteria were reassessed before proceeding with any
statistical comparison. In instances where transformation was unable to
produce data satisfying the conditions for parametric comparison, data were
instead subjected to non-parametric analysis using the Kruskal-Wallis test. If, by
applying the Kruskal-Wallis test to all the data indicated a significant variation
between discrete groups, individual pairwise comparisons were performed
using InVivoStat (British Association for Psychopharmacology, Cambridge, UK),
which recommends the Behrens-Fisher test in preference to the Mann-Witney
(although both tests generally gave equivalent results), primarily because it
allows for non-normal distributions (e.g. skewed) and heteroscedascity of the
data (Munzel & Hothorn, 2001). Furthermore, when comparing the analysis of
known data-sets, the Behrens-Fisher test gave the most robust statistics
(Steland et al., 2011).
Parametric one-way ANOVA and t-test comparisons were performed using
InStat 3 (GraphPad Software, La Jolla, California, USA), with all-pairwise
comparisons performed, and adjusted for multiple comparisons with the method
of Bonferroni. For the time-course analysis of light-dark box results, repeated
measures ANOVA was performed, with individual pairwise comparisons
performed for each time bin relative to bin 1, using Dunnett’s test.
Parametric two-way ANOVA were performed using InVivoStat to report the
Fisher’s least-significant difference (LSD) p-values from planned individual
pairwise comparisons (i.e. not every individual pairwise comparison was made,
Chapter 2: Materials and Methods 79
but only a selected proportion were chosen, on the basis of scientific relevance,
before the experiment commenced). Even though the probability of a type I
error (inappropriate rejection of a null hypothesis) was reduced by using only
planned comparisons, the risk of such an error is high when considering the
large numbers of comparisons made within each experiment (as high as 25).
We therefore adjusted the LSD p-values using the method of Benjamini and
Hochberg (Benjamini et al., 2001). This is a means to control the 'false
discovery rate', which limits the proportion of type I errors to 5%. When
considering such a large number of comparisons, we favour this approach over
that of Bonferroni (which reduces the significance level, alpha, in proportion to
the number of comparisons made). The Bonferroni method controls the family-
wise error rate, ensuring a 5% chance of a single type I error from all
comparisons made, but this increased robustness comes at the price of a high
type II error rate (failing to reject null hypotheses that should be rejected).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 80
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid
potentiation at the GABAA receptor α2 subunit
3.1. Introduction
As discussed in the Section 1.2.2, observations that GABAA receptor Cl-
conductance is modulated in a stereo-selective and biphasic manner by
neuroactive steroids led to proposals that there are two specific neurosteroid
binding sites on the GABAA receptor: an “activation site” and a “potentiation
site” (Harrison & Simmonds, 1984; Harrison et al., 1987; Puia et al., 1990; Paul
& Purdy, 1992). In defining these neurosteroid binding sites, Hosie et al. (2006)
demonstrated that hydrophobic substitution of a critical α1Q241 residue was
sufficient to selectively disrupt neurosteroid potentiation. For example, mutation
α1Q241M eliminated potentiation by allopregnanolone, with minimal effects on
receptor activation by GABA and modulation by other allosteric ligands
(barbiturates, benzodiazepines). This work also demonstrated that the full effect
of neurosteroid binding to the activation site is only achieved by concomitant
binding of neurosteroid molecules to the potentiation site (i.e. mutation α1Q241M
disrupts both potentiation and direct activation by neurosteroids). Hydrophobic
substitution of the conserved α1Q241 residue has been extended to the other five
α subunit isoforms, in every case the substitution disrupted potentiation by
neurosteroids (Hosie et al., 2009).
This discovery provided a unique opportunity to dissect the roles of the various
α subunit isoforms in the physiological and pharmacological functions of
neurosteroid molecules. By creating a knock-in transgenic mouse in which
genomic DNA encoding specific α subunit has been replaced with a Q-to-M
mutant copy of the receptor, one can assess the consequences of losing
neurosteroid modulation of that receptor subtype on synaptic inhibition and
mouse behaviour. Whilst such questions could be addressed with α-subunit
knock-out mice, this approach is often hampered by lethality (e.g. γ2 knock-out
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 81
mice are not viable, restricting study to heterozygous knock-outs (Gunther et al.,
1995)) or results in compensatory changes in expression of other subunits (e.g.
α1-/- mice show increased expression of α2 and α3 subunits (Kralic et al.,
2002)). With the knock-in approach, mutant receptors should be expressed and
respond to GABA as if wild-type (Hosie et al., 2006). Phenotypic changes in the
mouse would therefore probably result from lack of neurosteroid function at this
receptor, rather than a compensatory alteration in expression. However, caution
should be exercised when making such an assumption: our model would
involve losing response to endogenous molecules, and the effects of losing any
such modulation under baseline conditions may be significant, and theoretically
could result in compensatory changes. The latter half of this chapter deals with
verifying a lack of compensation in our transgenic model.
In this body of work, the Q-to-M substitution has been introduced in the α2
subunit – i.e. knock-in mice harbouring the mutation α2Q241M have been
generated. Whilst the α2Q241M mutation has been shown to disrupt potentiation
by neurosteroid (Hosie et al., 2009), further criteria must be met by the mutation
for it to be appropriate for the mouse model; specifically that GABA sensitivity is
preserved, and subunit expression should be unperturbed. We would also
expect benzodiazepine potentiation to be unaffected by the mutation. These
features have been verified in this study by characterising responses of α2β3γ2
GABAA receptors expressed in HEK293 cells. This approach was also used to
probe the effects of concurrent modulation of these receptors with neurosteroids
and benzodiazepines.
The α2 isoform was selected as the first candidate for generating a knock-in
mouse because of its strong links with the mammalian anxiety circuitry (see
Section 1.3.2), and we predict that neurosteroid anxiolysis will occur via α2-type
GABAA receptors. This hypothesis is addressed in later chapters, but firstly
screens must be performed to ensure a lack of underlying compensatory
changes that could account for any behavioural effects of the mutation. In this
study, assays to measure GABAA receptor expression at the protein level have
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 82
been chosen (Western blot and quantitative immunofluorescence), as these are
the proteins that ultimately mediate inhibitory transmission. GABAA receptor
subunits α1-α5 were selected as the most likely candidates to be up- or down-
regulated in response to the mutation, and so subjected to quantitation. Assays
were focussed on four brain areas that express GABAA receptor α2 subunits –
cortex, hippocampus, cerebellum and nucleus accumbens (Laurie et al., 1992a;
Wisden et al., 1992; Sperk et al., 1997; Pirker et al., 2000; Sieghart & Sperk,
2002).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 83
3.2. Results
3.2.1. Confirming that α2Q241M has no effect on GABA and diazepam sensitivity
of αβγ receptors
The effects of α2Q241M on receptor function were examined by characterising α2-
containing GABAA receptors expressed in HEK293 cells. This cell line has been
used widely as a platform for expression and study of GABAA receptor
properties. Because of their non-neuronal origin, these cells do not express
endogenous GABAA receptors (although there has been evidence for some low
level expression of β3, γ3 and ε subunits (Thomas & Smart, 2005)), allowing
one to control the composition of GABAA receptor heteropentamers under
study. For electrophysiology, cells were transfected with expression plasmids
for each of the subunits in the combination α2β3γ2S, together with an
expression plasmid for eGFP (see Section 2.5). This combination of receptor
subunits is thought to be representative of native α2-type GABAA receptors in
vivo (Benke et al., 1994; McKernan & Whiting, 1996). The eGFP allows
identification of transfected cells, by their green fluorescence, for study by
whole-cell patch clamp electrophysiology.
Peak current responses of wild-type (α2β3γ2S) and mutant (α2Q241Mβ3γ2S)
receptors were measured in response to brief (2-4 s) applications of GABA at
concentrations 0.01-1000 µM (Fig. 3.1 C). GABA concentration-response data
were collected for each cell, and fitted with the Hill equation (Equation 1,
Section 2.5.3). Mean Hill-fit parameters, EC50 (the concentration of GABA
eliciting a 50% maximal response) and n (Hill coefficient), for wild-type and
mutant receptor combinations are not different (Table 3.1). This is consistent
with previous observations that the α2Q241M mutation is without effect on the
GABA sensitivity of the α2β3γ2S receptor (Hosie et al., 2009). The lack of effect
of the α2Q241M mutation on the GABA sensitivity can also be appreciated by
noting the good superimposition of the mean GABA concentration-response
curves for these receptors (Fig. 3.1 A).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 84
To confirm that the α2Q241M mutation did not affect benzodiazepine modulation,
concentration-potentiation curves for diazepam potentiation of wild-type and
mutant receptors were constructed, as outlined in Section 2.5.3. Briefly, peak
current responses of a transfected cell to 0.5 µM GABA (EC15 response) co-
applied with varying concentrations of diazepam were normalised to peak
current responses to 0.5 µM GABA alone within that cell. As above,
concentration-potentiation curves were constructed for each cell, this time using
a modified Hill equation, to include a pedestal (Equation 2, Section 2.5.3). The
mean fit parameters, summarised in Table 3.1, are very similar. There is a small
but significant increase in the Hill coefficient for mutant vs. wild-type receptors,
but when mean data are plotted together (Fig. 3.1 B) the concentration-
potentiation curves are superimposed. The diazepam sensitivity of the α2β3γ2S
receptor combination is therefore unchanged by α2Q241M mutation, consistent
with previous work characterising this mutation on α1 (Hosie et al., 2006).
These data demonstrate that this substitution specifically abolishes potentiation
by neurosteroids.
Dose relationship
wild-type Q241M
GABA concentration-response
EC50 (µM) 3.0 ± 0.8 3.2 ± 0.4
n 1.2 ± 0.2 1.2 ± 0.1
Diazepam concentration-potentiation
Pmin 0.9 ± 0.04 1.0 ± 0.03
Pmax 2.9 ± 0.2 2.9 ± 0.3
EC50 (nM) 50 ± 16 38 ± 11
n 1.0 ± 0.1 1.7 ± 0.3*
Table 3.1 – Hill fit parameters for GABA and diazepam responses
Summary of the fits (mean ± s.e.m.) for GABA and diazepam concentration-response
curves. Pmax and Pmin correspond to the maximum and minimum points on the
diazepam potentiation curve. The Q241M mutation significantly increases the Hill
coefficient (n) for diazepam potentiation (* p < 0.05 vs. wild-type, Mann-Witney test).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 85
Figure 3.1 – Mutation α2Q241M has no effect on the GABA or diazepam sensitivity
of α2β3γ2S receptors expressed in HEK293 cells
A. and B. Concentration-response curves for wild-type (blue) and mutant (grey)
α2β3γ2S receptors expressed in HEK293 cells. Peak GABA responses are
normalised to the maximal response in each cell (A). Potentiation of GABA EC15
responses by diazepam (B) is expressed as the peak response to GABA plus
diazepam relative to the response to GABA alone. Each graph represents the mean
responses across 7-8 cells (error bars, s.e.m.; 3-8 data points per concentration).
C. Representative membrane currents on short applications of the indicated GABA
concentrations (µM) to wild-type (blue) and mutant (grey) receptors
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 86
3.2.2. Probing the effect of α2Q241M on responses to both THDOC and diazepam
When considering the effect of pharmacological potentiators of GABAA
receptors in vivo, one must not forget that neurosteroid molecules are produced
endogenously. This suggests that whilst the potentiating effect of diazepam on
α2Q241Mβ3γ2S receptors may be normal when receptors are assessed in
HEK293 cells, the level of diazepam potentiation may not be ‘wild-type like’
when taking place on a background of endogenously-synthesised neurosteroid
in vivo.
To assess the significance of this, transfected HEK293 cells were used to
examine the effects of co-applying the neurosteroid THDOC together with
diazepam, on the potentiation of an EC15 GABA response. A typical experiment
is depicted in Fig. 3.2 C; briefly, peak current responses to 0.5 µM GABA were
compared with those achieved by co-application of 0.5 µM GABA together with
either DMSO vehicle (V), 100 nM THDOC (T) alone, 500 nM diazepam (D)
alone, or both THDOC and diazepam (T + D) together. As can be seen in Fig.
3.2 A, in wild-type α2β3γ2S receptor-expressing cells there is no effect of
vehicle, and the effects of T and D are similar, and are approximately additive,
when both substances are applied together. Conversely, in mutant
α2Q241Mβ3γ2S receptor expressing cells (Fig. 3.2 B), there is no effect of
THDOC alone or when co-applied with diazepam (i.e. potentiation by T + D co-
application is no different to that for D alone).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 87
Figure 3.2 – Mutation α2Q241M specifically abolishes the potentiating effect of
THDOC on GABA EC15 responses of α2β3γ2S receptors expressed in HEK293
cells
A. and B. Bar charts outlining the degree of potentiation achieved by co-application of
vehicle (V), 100 nM THDOC (T), 500 nM diazepam (D) or both THDOC and diazepam
(T + D) with 0.5 µM GABA (G). Responses are normalised to the preceding peak
response to 0.5 µM GABA alone. Effects of THDOC and diazepam are additive in wild
type receptors (A, n=11 cells); THDOC has no effect on α2Q241M mutant receptors (B,
n=6 cells). Significant effects of treatments are highlighted: ** p < 0.01 effect of
substance vs. G alone; ++ p < 0.01 effect of T + D vs. response to T alone or D alone
(ANOVA, multiple comparisons corrected with Benjamini-Hochberg adjustment).
C. Representative membrane currents for wild-type (blue) and α2Q241M mutant (grey)
α2β3γ2S receptors expressed in HEK293 cells.
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 88
3.2.3. Generating an α2Q241M knock-in transgenic mouse
The process of generating the mouse strain is described more fully in Section
2.2.1; this stage of the project was performed by performed by Mike Lumb, a
molecular biology technician in the lab, in collaboration with GenOway (Lyon,
France). Briefly, wild-type exon 8 of α2 subunit genomic DNA in embryonic stem
(ES) cells was precisely replaced with exon 8 containing the desired base-pair
changes (CAA to ATG) for point mutation Q241M. ES cells that have
successfully undergone homologous recombination were used to generate a
transgenic mouse line with the α2Q241M mutation on a C57BL/6J background.
Homozygous (hom) knock-in mice are viable and fertile, and there are no overt
effects of the mutation on the general appearance of the mice. Animals were
maintained as heterozygous (het) breeding pairs, to allow comparison of hom
knock-ins with wild-type (wt) littermates. This is preferable over separately
maintaining wt and hom lines, where genetic drift between the two populations
could contribute to, or account for, any observed phenotypic differences (Bailey,
1982).
3.2.4. α2Q241M has no effect on the expression of GABAA receptor subunits α1-
α4 in Western blot assays
Total protein was isolated from four brain regions of interest (cortex, cerebellum,
hippocampus and nucleus accumbens (NAcc)) as described in Section 2.3.1.
Equal amounts of protein from six wt, six het and six hom males were subjected
to Western blot analysis for each of the four GABAA subunits α1, α2, α3, and α4
(where detectable) using subunit-selective antisera (subunit specificity of
antibodies was confirmed by blotting recombinant receptors – for details see
Section 2.3.2).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 89
Figure 3.3 – No change in expression of GABAA subunits α1-α4
Bar charts detail the relative expression levels of each GABAA receptor subunit in
protein samples from each brain area. There are no significant effects of genotype on
expression of any subunit in any brain area tested (p values in black indicate results of
parametric one-way ANOVA, those in green refer to non-parametric Kruskal-Wallis test
results). n.d. - not determined because no reliable signal was obtained with calibration
blots. Bottom: representative Western blots.
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 90
The blotting conditions were optimised by running a series of titrations to
determine the amount of total protein which, when blotted under the conditions
outlined in Section 2.3.2, produces a signal in the middle of the dynamic range
of the detection system. When working under such conditions, one maximises
the chances of detecting any up- or down-regulation of GABAA subunit
expression. As can be seen from results of the Western blot assay (Fig. 3.3),
there are no significant effects of genotype on subunit expression for any of the
detectable subunits in any of the brain areas assayed.
3.2.5. α2Q241M has no effect on the expression of GABAA receptor subunits α1-
α5 in quantitative immunofluorescence assays
Three mice of each genotype (wt, het, hom) were sacrificed with a lethal
overdose of sodium pentobarbital and brains were fixed by cardiac perfusion as
described in Section 2.4.1. Thin (40 µm) coronal sections were obtained from
frozen brain, and subjected to immunofluorescent staining and confocal z-stack
image acquisition. Representative example images for the three areas studied –
CA1 region of the hippocampus, the dentate gyrus (DG) and the NAcc – are
shown in Fig. 3.4. The CA1 images encompass the cell body layer and apical
dendrites, the DG images encompass the granule cell layer and molecular layer
of the medial blade, and NAcc images include both core and shell regions.
The staining for α1 in both hippocampal regions is similar, showing a mixture of
punctate and diffuse staining, with some cells showing much more intense
staining of cell bodies and dendrites than others. Such a distribution of α1 has
been seen by other investigators staining under similar conditions, using a
different antiserum (Benke et al., 1994; Mtchedlishvili et al., 2003). Staining for
α1 in the NAcc was rather different, showing a variable number of bright
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 91
punctae against a high background. This led to problems with quantification, as
discussed in more detail below.
Immunostaining for α2 was similar and strong in all three areas examined.
Punctate staining is seen throughout cell body and dendritic layers of both
regions of the hippocampus. Similar staining for α2 in the DG has been
previously published (Benke et al., 1994). Uniformly-distributed punctate
staining is also seen throughout the NAcc core and shell (as would be expected
from the work of Pirker et al. (2000)).
As expected from immunohistochemical work by other investigators (Sperk et
al., 1997), immunofluorescent staining for α3 is much less intense than α1 and
α2, showing fewer bright punctae against a very low background. The
expression of α3 was rather heterogeneous throughout the NAcc: some clusters
of cells showing very intense staining, whilst other areas have fluorescence at
background levels. There were no obvious structural correlations that could
explain this variation – neither a difference in core vs. shell regions of the NAcc,
nor a clear anterior-posterior pattern of staining between coronal slices. The
only common feature noted was a thin bright band of staining surrounding a
dark region which often correlated with the very edge of the shell (asterisks on
Fig. 3.4 C), but this was not true of every slice (c.f. wt, het and hom images in
Fig. 3.4 C, all representing expression in a similar position of the NAcc). To
ensure the mean fluorescence measure of expression accounts for this patchy
distribution, images of α3 in the NAcc were acquired with a lower power lens
(20x magnification vs. 63x in all other images) to encompass a larger field of
view.
The α4 subunit shows a stronger expression in molecular layer than the granule
cell layer of the DG, with an almost opposite distribution in CA1 (i.e. intense
diffuse staining in the cell bodies and proximal regions of dendrites, and less
intense staining in the dendritic region). It was also noted that there was an
anterior-posterior variation in α4 expression, with more intense staining of
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 92
coronal sections corresponding to more anterior slices. This subunit also
appears to be expressed well in the NAcc, as was expected from
immunohistochemical work (Pirker et al., 2000), and shows a mixture of diffuse
and punctate staining uniformly throughout the structure.
The α5 isoform shows strong diffuse staining of cell bodies in all three regions
studied. Staining in CA1 is very much like that for α4, except that no clear
anterior-posterior variation was noted. In both CA1 and DG images, α5 can also
be seen to have punctate staining uniformly throughout the dendritic region.
This pattern for hippocampal α5 staining is most consistent with that seen by
Prenosil et al. (2006). Similar strong expression of α5 in cell bodies is noted for
staining in the NAcc, with no obvious difference between core and shell regions.
Interestingly, Wisden et al. (1992) fail to detect α5 subunit expression in the
NAcc at the level of mRNA. However, Pirker et al. (2000) observed α5 subunit
expression in the NAcc using immunohistochemistry, which is described as
weak and diffuse; unfortunately an image of this staining is not presented in this
work, precluding a detailed comparison with our α5 staining pattern.
Figure 3.4 – Immunofluorescent localisation of GABAA subunits α1-5 in coronal
sections of hippocampus and nucleus accumbens (images on following pages)
A. Representative images for immunofluorescent staining in the CA1 region of the
hippocampus. Images were acquired with the 63x lens at 1x zoom, and encompass the
cell body layer and the apical dendrites. Scale bar, 30 µm.
B. Representative images for immunofluorescent staining in the dentate gyrus. Images
were acquired with the 63x lens at 1x zoom, and encompass the granule cell layer and
molecular layer of the medial blade. Scale bar, 30 µm.
C. Representative images for immunofluorescent staining in the nucleus accumbens.
Images for α1, α2, α4 and α5 were acquired with the 63x lens at 1x zoom; images for
α3 were acquired with the 20x lens at 1x zoom. Scale bars, 30 µm. Red asterisks
highlight examples of where strong α3 subunit immunoreactivity corresponds with the
outer edge of the nucleus accumbens shell.
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 93
Fig 3.4 A – Hippocampal CA1 immunofluorescence: representative examples
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 94
Fig 3.4 B – Dentate gyrus immunofluorescence: representative examples
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 95
Fig 3.4 C – Nucleus accumbens immunofluorescence: representative examples
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 96
When quantifying the immunofluorescence, the simplest possible approach was
employed to ensure that most of the image data was taken into account, and
that analysis bias was minimised. For images of DG and CA1, z-stacks were
split into two large regions of interest (ROIs) that encompass either the cell
body layer or the dendritic region. The mean fluorescence intensity of the entire
ROI was measured at three points in the z-stack (top, middle, bottom) and
averaged. This process was performed for images from three coronal sections
each from three animals of each genotype (i.e. mean values represent data
from 9 z-stacks). This approach was chosen, rather than attempting to quantify
expression in finer detail in individual cells, because there is no experimenter
bias that would be involved in defining cells or dendrites for quantitation. To
quantify expression in the NAcc, mean fluorescence intensity was measured for
the entire image, rather than a specific ROI. The results of this simple method of
quantitation are depicted in Fig. 3.5; no significant variation of subunit
expression with genotype was detected for any of the subunits tested.
On closer inspection, one may note that some subunits show a trend towards a
difference in subunit expression across genotypes: α3 expression tends to be
reduced in the hom DG cell bodies and dendrites; α4 and α5 tend toward
increased expression in hom DG dendrites; and α1 expression tends to be
lower in the hom NAcc. These cases are therefore considered in more detail.
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 97
Figure 3.5 – Quantitation of immunofluorescent staining for GABAA subunits α1-
α5 in coronal sections of hippocampus and nucleus accumbens
The relative mean fluorescence intensities for each subunit in each brain area are
summarised in these bar charts. See Section 2.4.3 for details. There are no significant
effects of genotype on expression of any subunit in any brain area (p values refer to
Kruskal-Wallis test results).
The contribution of background fluorescence to the mean fluorescence intensity
values measured here can be a source of error. For example, there are few α1
punctae against high background fluorescence in the NAcc, so the background
dominates the mean fluorescence intensity, and could account for the variation
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 98
seen in apparent expression. A similar problem is faced for α3 expression in the
DG. In these cases, images were examined with the FociPicker3D plug-in in
ImageJ, which identifies and quantifies immunopositive punctae by searching
for local fluorescence maxima in three-dimensions. With these images, the
plug-in is successful at identifying punctae against the background staining (see
Section 2.4.3 and Fig. 2.3). Results of this quantitation (Fig. 3.6) confirm that
there are no changes in the density of immunopositive punctae (punctae per
unit volume), the size of these punctae, or their intensity, across genotypes.
Interestingly, the number of α1-positive punctae was highly variable from slice-
to-slice for the NAcc (e.g. one het animal gave slices with 112, 234 and 3874
punctae). There was no obvious positional correlation, either anterior vs.
posterior or shell vs. core, that could explain this variability. Nevertheless, the
average puncta densities across all images show no clear variation with each
genotype, nor are those punctae any different in size or intensity between
genotypes.
The FociPicker3D approach was not successfully applied to the α4 and α5
images because they contain a mixture of punctate and diffuse staining (see
Section 2.4.3 and Fig. 2.3 for more details). Most α5 subunit staining is
concentrated in the cell body layer in the DG, so the tendency toward
fluorescence differences in the molecular layer is not a cause for concern. In
contrast, α4 in the DG is more strongly expressed in the molecular layer, so a
tendency toward increased mean fluorescence intensity in this region may
represent an altered expression of this subunit in the knock-in mice.
Statistically, however, this difference is not significant: the overall Kruskal-Wallis
test produces a p value of 0.23. Individual pairwise comparisons also fail to
reach significance (wt vs. het p=0.20, wt vs. hom p=0.20, het vs. hom p>0.99).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 99
Figure 3.6 – Quantitation of immunofluorescent punctae for GABAA subunits α3
in the dentate gyrus and α1 in the nucleus accumbens
The relative mean density, size and intensities of immunopositive punctae for α3 and
α1 subunits in the dentate gyrus (DG) and nucleus accumbens (NAcc), respectively, in
coronal sections from wt, het and hom animals. Punctae were defined using the
FociPicker3D plug-in for ImageJ software (see main text for details), using either entire
z-stack images, or just the dendritic ROI (for DG α3 expression). There are no
significant effects of genotype on expression of either subunit (p values refer to
Kruskal-Wallis test results).
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 100
3.3. Discussion
3.3.1. α2Q241M and GABAA receptor function
The mutation α2Q241M has already been shown to abolish potentiation by the
neurosteroid allopregnanolone (Hosie et al., 2009). Here we extend this finding
to the neurosteroid THDOC, which fails to potentiate EC15 GABA responses at
100 nM, when α2Q241M is expressed in the receptor combination α2β3γ2S (Fig.
3.2). GABA sensitivity is retained in α2Q241M-mutant receptors, which also show
similar maximal current responses (1 mM GABA responses: wt – 950 ± 100 pA,
Q241M – 1400 ± 180 pA; p=0.07, unpaired t-test), which is consistent with
unaltered cell surface expression of the mutant receptors in HEK293 cells.
Assuming that these properties extend to α2Q241M receptors when expressed in
vivo, transgenic α2Q241M knock-in mice would be preferred over α2 knock-out
mouse to study the roles of the GABAA receptor α2 subunit in physiological and
pharmacological responses to neurosteroids: in the knock-in, the retained
expression and GABAergic function of the α2-type GABAA receptor would
ensure that phenotypes of the mouse result only from the loss of neurosteroid
function at that receptor.
Although not probed in this study, it is expected from previous work (Hosie et
al., 2006), that neurosteroids will retain action at the direct activation site.
Nevertheless, the concentrations of neurosteroids required to achieve direct
activation of the GABAA receptor are in the micromolar range – current
measurements of neurosteroid levels suggest such concentrations will at best
be at the very extremes of physiology (e.g. neurosteroid measures in pregnancy
peak at 100 nM (Paul & Purdy, 1992)), and probably play little role in normal
function of neurosteroids. Direct activation may become relevant during
experiments probing responses to injected neurosteroid molecules, where
levels at the synapse may well climb to those required for direct activation. In
addition, the full effect of direct activation by neurosteroid appears to require
concomitant binding and function of neurosteroids at the potentiation site (Hosie
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 101
et al., 2006). Both potentiation and direct activation by neurosteroids will
therefore be defective at α2-type GABAA receptors in α2Q241M knock-in mice.
As was demonstrated for the equivalent mutation on α1-type GABAA receptors
(Hosie et al., 2006), α2Q241Mβ3γ2S receptors have a wild-type-like response to
diazepam potentiation (Fig. 3.1). This confirms that the mutation has not simply
disrupted the ability of the GABA receptor to be potentiated per se, but that the
lesion is specific to the action of neurosteroids via the potentiation site. The
1:1:1 ratio of cDNA expression vectors we used was sufficient in this study for
expression and co-assembly of the γ2 subunit with α2 and β3, as indicated by
their responses to diazepam (receptors lacking the γ2 subunit would be
diazepam-insensitive (Pritchett et al., 1989)).
3.3.2. No compensatory changes in GABAA receptor α subunit expression in
α2Q241M mice
Whilst it is generally suggested that a knock-in mouse is preferable over a
knock-out mouse, because there is a lower risk of compensatory up- or down-
regulation of expression of various proteins, this should always be confirmed.
For example, a microarray analysis of mice with an α1 or α2-subunit-directed
knock-in mutation in two residues (S270H, L277A) – that disrupt ethanol
sensitivity – reveal an array of changes to mRNA levels, implying multiple
changes to many genes (Harris et al., 2011). Furthermore, unlike the alcohol
(Harris et al., 2011) or benzodiazepine (Low et al., 2000) site models, our
transgenic model is losing sensitivity to a modulator that is present
endogenously. Any compensatory changes in protein expression, particularly
GABAA receptor subunits, resulting from knocking in the α2Q241M mutation would
be interesting, but would also confound the investigation: do any of the
phenotypes of the mouse result directly from loss of neurosteroid modulation of
GABAA receptor α2 subunits, or indirectly as a consequence of compensatory
alterations in the mouse?
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 102
There are several possible approaches in screening for compensatory changes.
We chose to focus on methods that assay expression at the protein level
(Western blot, immunofluorescence), rather than measuring mRNA levels
(quantitative PCR, microarray) because it is changes of protein expression
levels that are of functional importance. Our experiments have focussed solely
on the expression levels of GABAA subunits α1-α5 as likely candidates for
compensatory changes (the α6 subunit is not widely expressed, only being
present in cerebellar granule cells and the cochlear nucleus (Laurie et al.,
1992a; Pirker et al., 2000) and was not studied here). Analysing lysates of
cortex, hippocampus, cerebellum and NAcc for α1-α4 expression showed no
major global alterations in expression of these subunits. Not every subunit was
detectable in all brain areas by Western blot (Fig. 3.3), and attempts to utilise an
α5-selective antiserum in Western blots have proven unsuccessful, but this
subunit was successfully assayed by immunofluorescence. Similarly, the anti-α4
antiserum was unable to produce a reproducible, quantifiable, signal for
expression in whole-tissue lysates from hippocampus or nucleus accumbens,
but this subunit is nicely detected by immunofluorescence. Subunits α1-α3 were
successfully probed by both methods. Immunofluorescence allows subunit
expression patterns to be examined in finer detail than Western blotting.
Quantitative immunofluorescence focussed on subunits α1-α5 in the areas that
were subjected to electrophysiological analysis (CA1, DG and NAcc, see
Chapter 4). Overall, neither method has demonstrated any significant change in
subunit expression for any of the five subunits.
When comparing immunofluorescent staining obtained in this study with
immunostaining patterns in other published work, similarities and differences
can be noted. In accord with work by Mtchedlishvili et al. (2003) and Benke et
al. (1994), α1 antiserum stains some cells and processes in hippocampal slices
much more brightly than others, but is otherwise characterised by widespread
immunopositive punctae. The low abundance of α3 staining in the hippocampus
is also in agreement with observations by others (Sperk et al., 1997; Pirker et
al., 2000; Prenosil et al., 2006). Other commonalities include the more intense
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 103
staining for α2 and α4 subunits in the molecular layer of the dentate gyrus
(Benke et al., 1994; Pirker et al., 2000; Prenosil et al., 2006). Poor expression
for α1 subunits in the NAcc has also been demonstrated at mRNA (Sarviharju et
al., 2006) and protein (Pirker et al., 2000) levels. In addition, the expression
pattern for α2 in the DG resembles that in work published by others (Benke et
al., 1994).
Variable hippocampal expression patterns have been described for GABAA
receptor α5 subunits in the literature. Our work and that of Prenosil et al. (2006)
finds strong staining in cell bodies, whilst Sperk et al. (1997) and Pirker et al.
(2000) demonstrate an absence of α5 from the cell body layers. Cell-body
dominant staining for α5 subunits has also been observed in hippocampal
sections from young (P7) rats (Ramos et al., 2004), but these investigators then
observe a pattern reminiscent of Sperk et al. (1997) and Pirker et al. (2000) in
adult animals. Conversely, punctate staining of cell bodies have been observed
in hippocampal slices from one month old mice (Farisello et al., 2012). The
staining patterns obtained in this study were highly reproducible and consistent
from animal-to-animal and slice-to-slice. Perhaps this apparent variability
between studies is a consequence of differences in animal age and species,
and detection methods and antibodies used.
The expression of α subunit isoforms in the NAcc has been reported at the level
of mRNA (Wisden et al., 1992) and protein (Pirker et al., 2000). The low level of
staining we observe for α1 subunits is in agreement with these studies.
Furthermore, our staining patterns for α2, α4 and α5 subunits in the NAcc are
consistent with the description provided in the work by Pirker et al. (2000), who
note that all three subunits show diffuse labelling, with the strongest signal for
α2, and weakest for α5. The NAcc expression levels of α3 are described as low
by both studies (Wisden et al., 1992; Pirker et al., 2000), but a figure is not
available for a detailed comparison with our α3 staining pattern.
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 104
One important caveat of the approach used in this study is that they are
reporting total protein levels, not just that at the cell surface. If
immunofluorescence was performed without triton-x-100, only cell surface
subunit expression would be revealed. However, the thickness of the slices
used in this study (40 µm) demanded the use of triton-x-100 for antibody access
to below-surface regions of the slice. Whilst immunostaining neuronal cultures
may seem a suitable alternative approach to bypass the accessibility issue, the
subunit expression patterns in a cultured neuron will be further removed from
natural physiology than obtaining slices from fixed adult mouse brain. Another
means to address this issue would be to co-stain slices for other markers; for
example looking at the apposition of GABAA receptor punctae with a pre-
synaptic marker, such as GAD65 or GAD67, would indicate which GABAA
receptor punctae represent functional, synaptic GABAA receptors. Such an
experiment would also help to demonstrate whether normal numbers of
synaptic connections are formed in hom mice.
One must also remember that simply because there is no change in the amount
of GABAA receptor α subunit, this does not guarantee that their modulation by
other means is unaffected. Phosphorylation of GABAA receptors represents a
major means of modulation in vivo, which can also interact with neurosteroid
modulation (see Sections 1.1.3 and 1.2.3). Perhaps future experiments should
probe GABAA receptor phosphorylation states in lysates from the α2Q241M mice
(e.g. using a phospho-specific antibody in Western blot).
Taken overall, the Western and immunofluorescence data presented here
suggest that the phenotypes of this transgenic mouse are not a result of large
changes in protein expression of the GABAA subunits α1-α5. It would be
surprising if loss of neurosteroid function at α2-type GABAA receptors is not
accommodated for in some way, and so it may be sensible to consider using a
microarray or proteomic approach to screen for differences in expression
patterns in hom vs. wt mice. Any ‘hits’ from this screen could then identify
targets for further validation. Furthermore, we have not assessed the levels of
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 105
neurosteroid in the mouse strain, which may represent another avenue for
compensatory changes.
3.3.3. Endogenous neurosteroids may modulate the in vivo response to other
GABAA receptor potentiators
Retained sensitivity of the Q241M mutant receptors to potentiators such as
benzodiazepines (Fig. 3.1) and pentobarbital (Hosie et al., 2006) suggests that
these compounds could be useful for positive controls in this study – for
example if the anxiolytic response to neurosteroids is diminished in the α2Q241M
mice, is this a specific consequence of losing the neurosteroid potentiation site,
or is there some general defect in the anti-anxiety circuitry? If the mice retain
normal responses to diazepam or pentobarbital, then the latter hypothesis could
be discounted.
However, there are a few complications to note. Firstly, neurosteroids are
endogenous molecules, and so any modulator injected into a mouse will have
its effects on GABAA receptors on top of the baseline modulation by
endogenous neurosteroids. Secondly, certain benzodiazepines induce an
increase in endogenous synthesis of neurosteroids, and these have been
shown to underlie a component of the full in vivo response to these compounds
(e.g. midazolam’s anti-seizure activity is partly mediated by neurosteroid
synthesis (Dhir & Rogawski, 2012)). By co-applying neurosteroid with diazepam
in recombinantly expressed receptors (Fig. 3.2), it can be seen how the α2Q241M
mutation could affect responses to diazepam in vivo, where endogenous
neurosteroids are present. This effect must therefore be considered in
experiments in later chapters that use diazepam and pentobarbital as positive
controls in these mice.
Chapter 3: Creating a transgenic mouse with disrupted neurosteroid potentiation at the GABAA receptor α2 subunit 106
3.4. Conclusions
1. The α2Q241M mutation specifically abolishes potentiation of
heteropentameric αβγ GABAA receptor function by neurosteroids, whilst
activation by GABA and potentiation by diazepam are unaffected when
channels are studied in a heterologous expression system.
2. Transgenic knock-in mice have been generated, in which the wild-type
copy of α2 genomic DNA has been precisely replaced with a Q241M
point-mutated version of the receptor.
3. The α2Q241M mutation has no overt effects on the expression of GABAA
receptor subunits α1-α5 in the transgenic strain, whether present in the
heterozygous or homozygous state – i.e. phenotypes of the knock-in
mice can be attributed to the loss of neurosteroid potentiation (rather
than altered GABAA receptor subunit expression and distribution).
4. Neurosteroids are present endogenously, and so the mutation may alter
responses to other GABAA receptor potentiators in a whole animal.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 107
Chapter 4: Electrophysiological consequences of losing neurosteroid
modulation at GABAA receptor α2 subunits
4.1. Introduction
Neurosteroids can enhance tonic and phasic GABA transmission, with either
response expected to reduce neuronal excitability by hyperpolarising the neuron
and/or by shunting excitatory inputs. The roles of α2 subunits with regards to
these responses can be defined using homozygous α2Q241M mutant mice: any
deficiency in the neurosteroid response could be attributed to the lack of
neurosteroid potentiation at these receptors. To screen for any phenotypic
effect of the α2Q241M mutation, this investigation focussed on cells within the
hippocampus and nucleus accumbens, which are known to strongly express the
GABAA receptor α2 subunit (Fig 3.4; Sperk et al., 1997; Pirker et al., 2000;
Sieghart & Sperk, 2002; Mohler, 2006a).
4.1.1. GABAergic neurotransmission in the hippocampus
The hippocampus is implicated in many processes, including cognition,
emotion, and the formation of spatial memories (Fanselow & Dong, 2010).
Moreover, hippocampal pathologies are linked to epilepsy, anxiety disorders
and depression (de Lanerolle et al., 2003; Bannerman et al., 2004; MacQueen
& Frodl, 2011). The cytoarchitecture and circuitry of the hippocampus have
been well defined and comprise two interlocking layers of pyramidal and
granule cells (Fig. 4.1), both of which are under extensive GABAergic control.
Hippocampal inputs into the hypothalamus are a major site for stress activation
of the HPA axis (Brown et al., 1999) and the hippocampus is proposed to be the
site at which neurosteroids limit HPA axis activation by enhancing GABAergic
inhibition in this area (see Section 1.3.1). Some of the anxiolytic function of
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 108
endogenous neurosteroids may involve their action in the hippocampus, and we
propose that α2-type GABAA receptors will play a central role in this response.
We have therefore examined inhibitory neurotransmission within the
hippocampus of α2Q241M knock-in mice using whole-cell patch clamp
electrophysiology. Experiments have focused on CA1 pyramidal cells (CA1
PCs) and dentate gyrus granule cells (DG GCs) within acute brain slices.
Effects of the α2Q241M mutation are likely to be detected in these cell types, as
both express the α2 subunit. Furthermore, the enrichment of α2 subunits at the
axon initial segment of CA1 PCs places this receptor in an ideal position to
inhibit action potential generation (Nusser et al., 1996) and this subunit has
already been demonstrated to play a significant role in generating IPSCs in
these cells (Prenosil et al., 2006). Neurosteroid-mediated potentiation of these
receptors may therefore have profound consequences for firing of these
neurons within the hippocampal circuit.
A number of observations suggest that tonic currents may dominate in
mediating the anxiolytic response to endogenous neurosteroid. The increased
neurosteroid efficacy at δ-containing receptors makes them better poised to
respond to low levels of endogenous neurosteroids (Belelli et al., 2002; Brown
et al., 2002; Stell et al., 2003). Furthermore, the anxiolytic response to
neurosteroid is impaired in δ-/- mice (Mihalek et al., 1999), and Shen et al.
(2007) find that reduced tonic transmission in the hippocampus induces anxiety
during puberty in female mice. This may lead to the assumption that α2
subunits, which are thought to traffic to synaptic, as opposed to extrasynaptic,
sites (Nusser et al., 1996; Prenosil et al., 2006), will play little role in
neurosteroid-mediated anxiolysis. However, we are unaware of any evidence
directly excluding α2 subunits from passing tonic currents. Indeed, synaptic
receptors could be recruited to support tonic inhibition under certain conditions
(e.g. increased ambient GABA and/or endogenous neurosteroid levels in the
synapse could tonically open these channels). Furthermore, receptors can
move into and out of synaptic sites by lateral mobility within the membrane
(Thomas et al., 2005); when present peri- or extra-synaptically, classically
‘synaptic’ α2-type receptors could be involved in passing tonic currents, if they
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 109
are activated by this exposure to ambient GABA. Indeed α5βnγ2 receptors have
such a dual role, being responsible not only for CA1 PC tonic currents (see
Section 1.1.1), but also for a proportion of IPSCs within these cells (‘GABAslow’
IPSCs (Banks et al., 1998; Prenosil et al., 2006)). Finally, roles for synaptic
transmission in anxiety cannot be ruled out: synaptic α2-containing receptors
mediate benzodiazepine anxiolysis (Low et al., 2000), and several investigators
find hippocampal IPSCs are modulated by physiologically-relevant
concentrations of neurosteroid (e.g. Harney et al., 2003). The contribution of α2-
type GABAA receptors to both types of GABA current has therefore been
assessed in our investigation.
Figure 4.1 – Hippocampal networks relevant to anxiety and depression
The hippocampus has two major subdivisions, the cornu ammonis (CA) and the
dentate gyrus (DG). Both components can be divided into ‘cell body’ and ‘molecular’
layers. Principal cells reside in the cell body layers: in the CA region, these are
pyramidal cells (blue line) that can be divided into three regions (CA1-CA3), and in the
DG, principal cells are granule cells (green line). Inhibitory GABAergic interneurons
(not shown) are found in the molecular layers. Highlighted is the classical ‘trisynaptic
circuit’ (black arrows), starting with inputs via the entorhinal cortex (EC), and ending at
the subiculum (sub). Red arrows indicate some of the connections to other brain
regions that are relevant to emotional state and mood disorders (Fanselow & Dong,
2010). Note that the connections are simplified, with many more reciprocal
connections existing between the various subregions of the hippocampus.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 110
4.1.2. GABAergic neurotransmission in the nucleus accumbens
The nucleus accumbens (NAcc) forms part of the mesolimbic dopamine
pathway, which is believed to play roles in addiction and depression (see
Section 1.4.3). NAcc activity is under extensive inhibitory control: the vast
majority (95%) of neurons in the NAcc are GABAergic medium spiny neurons
(MSNs), which not only project to outputs (such as the ventral palladium), but
also make collateral connections with one another (Heimer et al., 1997).
Reduced neurosteroid and GABAergic function is implicated in the aetio-
pathology of depression. We propose that the antidepressant functionality of
neurosteroids could involve potentiation at α2-type GABAA receptors in the
NAcc.
Synaptic currents in the NAcc have been shown to be, at least in part, mediated
by α2-type GABAA receptors: Dixon et al. (2010) find mIPSCs are smaller and
faster decaying in α2-/- mice, without any compensatory change in α subunit
expression that could account for this change in IPSC properties. Furthermore,
these investigators demonstrated that repeated stimulation of signalling through
α2-type GABAA receptors (using Ro15-4513 in α2H101R knock-in mice) induces
behavioural sensitisation to Ro15-4513 and to cocaine, in a mechanism
independent of dopamine release in the NAcc (Morris et al., 2008; Dixon et al.,
2010). Thus, although signalling through α2-type GABAA receptors may not be
directly involved in the rewarding effects of addictive drugs (Dixon et al., 2010),
synaptic signalling through this receptor does appear to be involved in
strengthening behaviour toward the cues associated with drug taking (another
aspect of addiction). By extension, it is possible that the same signalling is
involved in learning to direct behaviour toward natural rewards, and so
dysfunctional phasic transmission through α2-type GABAA receptors could
contribute to depression. We therefore decided to characterise inhibitory
neurotransmission in NAcc MSNs, using α2Q241M mice to elucidate any roles for
neurosteroid potentiation at α2 subunits in modulating both tonic and phasic
transmission.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 111
4.1.3. Modulation of synaptic and tonic GABAergic currents by benzodiazepines
and neurosteroids
Spontaneous synaptic events (sIPSCs) and tonic currents have been assessed
in CA1 PCs, DG GCs and NAcc MSNs within acute brain slices from our
transgenic mouse strain. Properties of individual sIPSCs were assessed by
fitting procedures to define the synaptic event rise time, amplitude, weighted
decay time ( w) and charge transfer (area). The frequency of sIPSCs was also
monitored by measuring the inter-event interval (I.E.I.). The level of tonic
currents was measured by determining the change in root mean square (r.m.s.)
current noise on application of 20 µM bicuculline. Baseline parameters were
compared across genotypes to define any effects of the α2Q241M mutation on
basal inhibitory neurotransmission, both phasic and tonic.
Following stable baseline recording, GABAA receptor modulators THDOC or
diazepam were applied. These modulators are expected to increase the
duration of the decay phase of IPSCs: benzodiazepines (Otis & Mody, 1992;
Bai et al., 2001; Nusser & Mody, 2002; Dixon et al., 2008), and neurosteroids
and their analogues (Belelli & Herd, 2003; Harney et al., 2003) have been
shown to increase the decay times of miniature IPSCs (mIPSCs) from CA1
PCs, DG GCs and NAcc MSNs in acute brain slices. Reported effects on IPSC
amplitude vary – sometimes an increase is observed (e.g. Prenosil et al., 2006)
– but often amplitude is unchanged (e.g. Otis & Mody, 1992); no increase in
amplitude would indicate that synaptic receptors are saturated by the GABA
released during baseline transmission. Any effects of these potentiators on
IPSC frequency would indicate a pre-synaptic action, which has been observed
in some, but not all neural circuits (e.g. allopregnanolone increased the
frequency of GABAergic events in Xenopus laevis motorneurones (Reith &
Sillar, 1997), but THDOC has no effect sIPSC frequency recorded from GCs in
the cerebellum or DG (Stell et al., 2003)). The sensitivity of tonic currents to
neurosteroids and diazepam depend on the subunit composition of GABAA
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 112
receptors passing the currents: responses to classical benzodiazepines require
the presence of a γ subunit and absence of the benzodiazepine-insensitive α4
or α6 isoforms (Pritchett et al., 1989; Wieland et al., 1992); and although
neurosteroids potentiate via the α subunit (i.e. will potentiate all GABAA receptor
currents), the δ subunit confers a greater efficacy to neurosteroid potentiation
than γ-containing equivalents (Belelli et al., 2002; Brown et al., 2002; Hosie et
al., 2009). Here we have assessed the effects of knocking-in α2Q241M on the
modulation of synaptic and tonic GABAergic currents by benzodiazepines and
neurosteroids.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 113
4.2. Results
4.2.1. α2Q241M alters baseline inhibitory neurotransmission in acute slices of
hippocampus and nucleus accumbens
All electrophysiological experiments started with a period of stable baseline
recording, and data from these control epochs can be compared across
genotypes – any genotypic differences will indicate baseline effects of the
mutation. Because this mutation has no effect on GABA sensitivity (Section
3.2.1; Hosie et al., 2006; Hosie et al., 2009), nor an effect on GABAA receptor
expression (subunits α1-α5, Sections 3.2.4, 3.2.5), one may predict that
baseline inhibitory neurotransmission will be unaltered in knock-in mice
compared to wt littermates. However, endogenous neurosteroids may play a
significant role in modulating inhibitory neurotransmission recorded from brain
slice tissue (Belelli & Herd, 2003; Puia et al., 2003); if the α2 subunit is involved
in responding to these endogenous steroids, genotypic differences in phasic or
tonic GABA transmission will be expected at baseline. We have recorded from
MSNs in the two main anatomical regions of the NAcc – ‘shell’ and ‘core’ –
which are associated with distinct outputs, and functions (Heimer et al., 1997;
Shirayama & Chaki, 2006). However, we found no core vs. shell differences in
baseline transmission (nor responses to THDOC or diazepam), so these data
have been combined in the analyses below.
Phasic inhibition
Loss of neurosteroid potentiation at α2 subunits has some significant effects on
baseline inhibitory synaptic neurotransmission measured in all three cell types
(Table 4.1), most consistently a 20-30% decrease in w decay time for
recordings from homozygous animals compared to wild-types (Fig. 4.2). The
α2Q241M mutation also has a significant effect on w when present in the
heterozygous state for CA1 PCs, and tends toward an effect in het NAcc MSNs,
although this difference is not quite significant (NAcc MSN, wt vs. het; p =
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 114
0.062, Behrens-Fisher test). Frequencies (measured by inter-event intervals)
and amplitudes of sIPSCs are unaffected by genotype in recordings from any of
the cell types. There is a slight tendency for sIPSC rise times to be faster in
knock-in animals, although the only significant reductions are for hom vs. wt
CA1 PCs and het vs. wt NAcc MSNs. The charge transfer per sIPSC (area
values in Table 4.1) correlates well with w, tending to be decreased in
recordings from knock-ins vs. wild-types. Reductions in sIPSC area vs. wild-
types are significant for het CA1 PCs and NAcc MSNs, and for hom DG GCs.
Figure 4.2 – α2Q241M speeds the decay of baseline IPSCs
Representative baseline sIPSCs recorded from CA1 PCs, DG GCs and NAcc MSNs of
each genotype (each event represents an average of at least 100 individual IPSCs).
The events have been displayed with a peak-scaled amplitude to allow comparison of
IPSC decays. Events in cells from wt animals (black) are slower to decay than those
from het (blue) and hom (red) animals.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 115
Baseline sIPSC Parameters wt het hom
CA1 PCs
(25 wt, 15 het, 15
hom)
rise time (ms) 1.85 ± 0.09 1.59 ± 0.09 1.49 ± 0.10*
amplitude (pA) -31.6 ± 1.4 -31.9 ± 1.5 -36.5 ± 1.9
w (ms) 15.2 ± 0.5 11.8 ± 0.4*** 11.5 ± 0.6***
area (pA.ms) -383 ± 19 -304 ± 14** -348 ± 26
I.E.I. (ms) 202 ± 27 146 ± 37 145 ± 27
DG GCs
(19 wt, 21 hom)
rise time (ms) 2.02 ± 0.13 - 1.84 ± 0.13
amplitude (pA) -39.9 ± 2.3 - -37.6 ± 1.6
w (ms) 19.0 ± 1.0 - 13.7 ± 0.7****
area (pA.ms) -654 ± 69 - -424 ± 26***
I.E.I. (ms) 595 ± 134 - 401 ± 44
NAcc MSNs
(32 wt, 45 het, 20
hom)
rise time (ms) 2.45 ± 0.09 2.13 ± 0.07* 2.43 ± 0.19
amplitude (pA) -46.8 ± 2.3 -42.9 ± 2.0 -45.8 ± 3.4
w (ms) 26.5 ± 1.6 22.8 ± 1.2 21.0 ± 1.7*
area (pA.ms) -1050 ± 74 -845 ± 70* 857 ± 101
I.E.I. (ms) 478 ± 52 613 ± 64 744 ± 160
Table 4.1 – Comparing baseline sIPSCs recorded from acute brain slices
sIPSC properties (mean ± s.e.m.) from control epochs in the three cell types. In this
table, and those that follow, numbers in brackets indicate number of cells (recordings in
each case have been obtained from at least 4 animals). Statistically significant
differences are highlighted: * - p<0.05, ** - p<0.01, *** - p<0.001, and **** - p<0.0001 vs.
wt. (all pairwise Behrens-Fisher comparisons for CA1 PCs and NAcc MSNs, unpaired t-
tests for DG GCs).
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 116
Tonic inhibition
The baseline tonic currents recorded from all three cell types showed no
consistent variation with genotype (Table 4.2). NAcc MSNs tend toward
reduced amplitude with the α2Q241M mutation (i.e. tonic current wt>het>hom),
but there is no overall effect of genotype on tonic current (one-way ANOVA for
effect of genotype, p=0.256). Genotype also has no effect on the tonic current
observed in CA1 PCs (one-way ANOVA for effect of genotype, p=0.341), or DG
GCs (unpaired t-test wt vs. hom, p=0.883).
Baseline tonic currents wt het hom
CA1 PCs (20 wt, 14 het, 13 hom) 0.73 ± 0.11 0.80 ± 0.16 0.81 ± 0.16
DG GCs (19 wt, 21 hom) 0.31 ± 0.06 - 0.33 ± 0.09
NAcc MSN (18 wt, 18 het,17 hom) 0.61 ± 0.08 0.49 ± 0.07 0.40 ± 0.07
Table 4.2 – Comparing baseline tonic currents in slice recordings
Tonic currents (mean ± s.e.m.) measured in CA1 PCs, DG GCs and NAcc MSNs.
Numbers represent the root mean square (r.m.s.) current noise (pA) attributable to
GABAA receptor currents (i.e. the difference between the r.m.s. noise in control epochs
and that recorded after bicuculline treatment). Numbers in brackets indicate number of
cells (recordings in each case were obtained from at least 4 animals).
4.2.2. α2Q241M specifically diminishes neurosteroid potentiation of synaptic
inhibitory neurotransmission
Following stable baseline recording, GABAA receptor modulators THDOC or
diazepam were applied. Effects of these compounds on sIPSC parameters are
expressed as a percentage change from the control condition within that cell. To
confirm that any observed effects were due to GABAA receptor potentiation,
rather than, for example, an effect of the DMSO vehicle or the prolonged
recording period, ‘mock’ experiments were performed (see Section 2.6.2). If α2
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 117
subunits are significantly involved in the neurosteroid response of these
currents, the response to THDOC will be reduced in slices from knock-in
animals. Because the sensitivity to diazepam should be unaffected by the
α2Q241M mutation (Section 3.2.1), diazepam responses serve as a positive
control for GABAA receptor potentiation, and any deficit in response may
indicate some compensatory change is present in the knock-in mouse.
THDOC treatment of CA1 PCs, as expected, increased sIPSC decay time (Fig.
4.3 and Table 4.3). In cells from all three genotypes, 100 nM THDOC has
significant effects compared to ‘mock’-treated cells, but effects of the 50 nM
application were only significant for CA1 PCs from wt and het animals (for hom
animals, 50 nM THDOC vs. mock, p=0.693). The extent of w prolongation also
appears to depend on genotype: the 100 nM THDOC response of cells from
hom mice is significantly diminished compared to those from wt mice. The
response of cells from het animals to 100 nM THDOC is similar in magnitude to
wt, but there is a higher variance on this response, and data are not significantly
different to responses recorded in either wt or hom CA1 PCs.
The pattern of THDOC concentration-effects on sIPSC area mirrored the
changes seen for w (Table 4.3) – i.e. THDOC tends to increase the charge
transfer per event, and the effect appears smaller in cells from hom animals
than het or wt. However, there is a higher variance on these data, and only
overall effects of drug reach significance in two-way ANOVA analyses (effect of
drug, p<0.001, F=9.97, 2 degrees of freedom (d.f.); effect of genotype, p=0.249,
F=1.43, 2 d.f.; interaction p=0.605, F=0.69, 4 d.f.). None of the individual
pairwise comparisons between groups are statistically significant after
correction for multiple comparisons. There were no significant changes in rise-
times in response to THDOC, and no clear trend for a change in the sIPSC
amplitude with application of THDOC. There were no significant changes in
event frequencies in response to application of THDOC to CA1 PCs from
animals of any genotype.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 118
Figure 4.3 – THDOC effects on sIPSC decay time in CA1 pyramidal cells are
diminished in recordings from homozygous knock-ins
A. Bar chart detailing concentration-response relationships for THDOC effects on w
decay time for sIPSCs recorded from CA1 pyramidal cells in acute hippocampal slices
from animals of each genotype. ** p<0.01 and *** - p<0.001 for effect of THDOC vs.
mock, # - p<0.05 for wt vs. hom (all pairwise Behrens-Fisher comparisons).
B. Representative sIPSCs for CA1 PCs from wild-types and homozygotes. The
averaged control sIPSC (black) is superimposed on the averaged sIPSC after
equilibration with 100 nM THDOC (red). Each event represents an average of at least
100 individual IPSCs, and is displayed with amplitude peak-scaled to allow
comparison of IPSC decays.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 119
CA1 PC sIPSCs: drug effects wt het hom
rise time
mock -2.7 ± 4.7 (n=7) 2.3 ± 5.3 (n=4) -2.0 ± 4.1 (n=4)
50 nM THDOC 8.3 ± 6.9 (n=6) 3.8 ± 4.5 (n=5) 3.2 ± 4.3 (n=6)
100 nM THDOC 8.1 ± 4.1 (n=12) 9.8 ± 1.9 (n=6) 0.5 ± 3.1 (n=5)
amplitude
mock -2.8 ± 4.3 (n=7) -4.8 ± 8.7 (n=4) -8.1 ± 5.8 (n=4)
50 nM THDOC 18.3 ± 5.7 (n=6) -1.3 ± 3.0 (n=5) 1.2 ± 2.8 (n=6)
100 nM THDOC 6.3 ± 7.6 (n=12) 10.9 ± 7.5 (n=6) 4.8 ± 6.6 (n=5)
w
mock 6.4 ± 1.8 (n=7) 2.5 ± 1.6 (n=4) 5.7 ± 1.4 (n=4)
50 nM THDOC 18.5 ± 2.2 (n=6)** 18.6 ± 3.7 (n=5)*** 13.6 ± 3.7 (n=6)
100 nM THDOC 30.4 ± 3.0 (n=12)*** 32.9 ± 8.2 (n=6)*** 16.0 ± 1.0 (n=5)*** #
area
mock 4.1 ± 4.5 (n=7) -0.8 ± 10.6 (n=4) -2.9 ± 7.2 (n=4)
50 nM THDOC 40.5 ± 9.0 (n=6) 19.1 ± 5.2 (n=5) 16.1 ± 3.7 (n=6)
100 nM THDOC 37.2 ± 9.6 (n=12) 50.9 ± 18.6 (n=6) 22.8 ± 8.9 (n=5)
I.E.I
mock 13.3 ± 7.4 (n=7) -20.1 ± 13.2 (n=4) 0.74 ± 27.9 (n=4)
50 nM THDOC 80.0 ± 57.7 (n=6) 15.4 ± 8.0 (n=5) 8.2 ± 3.4 (n=6)
100 nM THDOC 22.7 ± 6.5 (n=12) 28.2 ± 31.5 (n=6) 6.7 ± 5.9 (n=5)
Table 4.3 – THDOC-induced changes CA1 PC sIPSC parameters
Percentage changes (mean ± s.e.m.) in CA1 PC sIPSC parameters in response to
THDOC application. ** p<0.01 and *** - p<0.001 for effect of THDOC vs. mock, # -
p<0.05 for wt vs. hom (all pairwise Behrens-Fisher comparisons).
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 120
As expected, the α2Q241M mutation diminished the sensitivity of synaptic
receptors to the neurosteroid THDOC in recordings from CA1 PCs. The impact
of genotype was apparent at the higher concentration of the neurosteroid, so
experiments in DG GCs and NAcc MSNs focused only on 100 nM THDOC.
These experiments also included diazepam as a positive control. The
concentration of diazepam chosen, 500 nM, induces a similar level of
potentiation as 100 nM THDOC when applied to recombinant GABAA receptors
expressed in HEK293 cells (Fig. 3.2). In slices from wt animals, both
substances have similar significant effects on sIPSC decay times in recordings
from DG GCs and NAcc MSNs (Fig. 4.4, Table 4.4, Table 4.5). Consistent with
our observations in CA1 PCs, the decay time prolongation by 100 nM THDOC
is significantly reduced in DG GCs and NAcc MSNs from hom animals. In
contrast, diazepam prolongation of sIPSC decay time is unaltered in DG GCs
and NAcc MSNs from the homozygous knock-in animals compared to those of
wt animals. These data are therefore consistent with the effects of the α2Q241M
mutation examined in HEK293 cells – disrupting neurosteroid potentiation
without any effect on benzodiazepine sensitivity.
As with CA1 PCs, 100 nM THDOC and genotype have no significant effects on
rise time, amplitude, area or frequency of sIPSCs recorded from DG GCs; there
is also no significant effect of 500 nM diazepam on these parameters (Table
4.4). Two-way ANOVA analysis of the area data reveal a significant effect of
treatment (p=0.034, F=3.75, 2 d.f.), but no overall effect of genotype (p=0.307,
F=1.08, 1 d.f.), and no interaction effect (p=0.207, F=1.65, 2.d.f.). None of the
individual pairwise comparisons survive correction for multiple comparisons, but
some of the unadjusted least significant difference (LSD) comparisons are of
note (wt mock vs. wt THDOC, p=0.035; wt mock vs. wt diazepam, p=0.023; and
THDOC wt vs. hom, p=0.056). It therefore seems that there are at least some
tendencies for the changes in charge transfer to correlate with changes seen
with w (i.e. THDOC being effective in wt/ineffective in hom, and diazepam
being effective in both wt and hom).
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 121
Figure 4.4 – Homozygous α2Q241M specifically disrupts neurosteroid
potentiation of sIPSCs in DG GCs and NAcc MSNs
A. THDOC and diazepam effects on w decay time for sIPSCs recorded from DG
GCs. ** - p<0.01 for effect of THDOC or diazepam vs. mock, + - p<0.05 for
diazepam vs. THDOC, ## - p<0.01 for wt vs. hom (ANOVA, multiple comparisons
corrected with Benjamini-Hochberg adjustment).
B. THDOC and diazepam effects on w decay time for sIPSCs recorded from NAcc
MSNs. ** p<0.01 for effect of THDOC or diazepam vs. mock treatment; + - p<0.05
and ++ - p<0.01 for effect of diazepam vs. THDOC; ## - p<0.01 for hom vs. het and
hom vs. wt (ANOVA, multiple comparisons corrected with Benjamini-Hochberg
adjustment).
C. and D. Representative sIPSCs for wild-type and homozygote DG GCs (C) and
NAcc MSNs (D). The averaged sIPSC in the control epoch (black) is superimposed
on the averaged sIPSC after equilibration with 100 nM THDOC (red) or 500 nM
diazepam (green). Each event represents an average of at least 100 individual
sIPSCs, and is displayed with amplitude peak-scaled to allow comparison of
decays.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 122
DG GC sIPSCs: drug effects wt hom
rise time
Mock 4.5 ± 1.8 (n=5) 13.3 ± 3.1 (n=7)
100 nM THDOC 2.8 ± 7.9 (n=7) 7.4 ± 7.6 (n=7)
500 nM diazepam 7.6 ± 4.5 (n=7) 21.7 ± 4.0 (n=7)
amplitude
Mock 17.3 ± 3.2 (n=5) 21.7 ± 8.8 (n=7)
100 nM THDOC 25.2 ± 7.8 (n=7) 16.5 ± 3.9 (n=7)
500 nM diazepam 30.0 ± 6.4 (n=7) 25.1 ± 7.1 (n=7)
w
Mock 14.6 ± 3.6 (n=5) 19.4 ± 5.2 (n=7)
100 nM THDOC 63.6 ± 7.0 (n=7)** 27.8 ± 5.7 (n=7) ##
500 nM diazepam 56.5 ± 5.2 (n=7)** 47.4 ± 5.7 (n=7)** +
area
Mock 42.2 ± 8.0 (n=5) 60.2 ± 18.9 (n=7)
100 nM THDOC 108.4 ± 21.2 (n=7) 53.8 ± 12.3 (n=7)
500 nM diazepam 113.9 ± 15.1 (n=7) 105.4 ± 30.6 (n=7)
I.E.I
Mock -14.2 ± 8.2 (n=5) - 3.9 ± 11.1 (n=7)
100 nM THDOC 19.0 ± 15.3 (n=7) -7.4 ± 11.3 (n=7)
500 nM diazepam -16.3 ± 5.4 (n=7) -12.0 ± 6.5 (n=7)
Table 4.4 – Changes in DG GC sIPSC parameters
Percentage changes (mean ± s.e.m.) in DG GC sIPSC parameters after application
of THDOC or diazepam. ** - p<0.01 for effect of THDOC or diazepam vs. mock, + -
p<0.05 for diazepam vs. THDOC, ## - p<0.01 for wt vs. hom (ANOVA, multiple
comparisons corrected with Benjamini-Hochberg adjustment).
In recordings from NAcc MSNs, there are no effects of drug or genotype on the
rise time, amplitude or frequency of sIPSCs (Table 4.5). Unlike the hippocampal
recordings, however, the effects on charge transfer reach statistical
significance: diazepam and THDOC significantly increase the areas of sIPSCs
recorded from wt and het animals, whilst only diazepam is effective in cells from
hom animals (i.e. no effect of THDOC in NAcc MSNs from hom mice).
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 123
NAcc MSN sIPSCs: drug effects wt het hom
rise time
mock 9.3 ± 5.0 (n=12) 6.6 ± 3.9 (n=15) 7.4 ± 5.3 (n=7)
100 nM THDOC 4.0 ± 4.2 (n=11) 5.6 ± 4.1 (n=21) 12.1 ± 11.0 (n=8)
500 nM diazepam 11.7 ± 10.8 (n=9) 14.8 ± 8.8 (n=9) 14.5 ± 5.1 (n=5)
amplitude
mock -2.9 ± 8.2 (n=12) -6.1 ± 5.9 (n=15) 7.7 ± 3.8 (n=7)
100 nM THDOC -6.2 ± 13.0 (n=11) 3.4 ± 2.9 (n=21) -2.3 ± 4.5 (n=8)
500 nM diazepam 10.9 ± 6.1 (n=9) 7.5 ± 9.9 (n=9) 12.3 ± 9.7 (n=5)
w
mock 8.1 ± 2.0 (n=12) 5.1 ± 2.0 (n=15) 7.2 ± 3.6 (n=7)
100 nM THDOC 35.1 ± 4.9 (n=11)** 41.8 ± 5.9 (n=21)** 14.4 ± 3.3 (n=8)##
500 nM diazepam 39.7 ± 5.2 (n=9)** 60.2 ± 9.8 (n=9)** +
41.7 ± 5.5 (n=5)** ++
area
mock 4.9 ± 9.2 (n=12) -2.9 ± 6.2 (n=15) 16.8 ± 8.3 (n=7)
100 nM THDOC 29.9 ± 26.9 (n=11)* 46.0 ± 6.6 (n=21)*** 14.7 ± 7.2 (n=8)
500 nM diazepam 52.2 ± 6.1 (n=9)* 73.2 ± 17.6 (n=9)* 60.3 ± 11.8 (n=5)** ++
I.E.I
mock 24.5 ± 9.4 (n=12) 35.2 ± 20.9 (n=15) 13.6 ± 14.0 (n=7)
100 nM THDOC 26.9 ± 12.1 (n=11) 16.7 ± 9.2 (n=21) 37.4 ± 34.7 (n=8)
500 nM diazepam 4.3 ± 9.6 (n=9) 12.4 ± 11.5 (n=9) -1.1 ± 11.5 (n=5)
Table 4.5 – Changes in NAcc MSN sIPSC parameters
Percentage changes (mean ± s.e.m.) in sIPSC parameters of NAcc MSNs after
application of THDOC or diazepam. * p<0.05, ** p<0.01 and *** p<0.001 for effect of
drug vs. mock treatment; + - p<0.05 and ++ - p<0.01 for effect of diazepam vs. THDOC;
## - p<0.01 for hom vs. het and hom vs. wt (ANOVA, multiple comparisons corrected
with Benjamini-Hochberg adjustment ( w) or all pairwise Behrens-Fisher comparisons
(area)).
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 124
4.2.3. α2Q241M specifically diminishes neurosteroid potentiation of tonic inhibitory
neurotransmission in dentate gyrus granule cells
DG GCs exhibited very little baseline tonic current in slices isolated from
animals of either genotype (Table 4.2). Nevertheless, application of 100 nM
THDOC significantly increased the r.m.s. current noise recorded from DG GCs
of wt mice (Fig. 4.5, Table 4.6). This increase in noise is of GABAA receptor
origin, because it is reversed by application of 20 µM bicuculline (Fig. 4.5 B).
This effect of THDOC is not apparent for DG GCs in slices from hom animals,
suggesting that it involves neurosteroid potentiation at α2-subunit-containing
receptors. The lack of effect of 500 nM diazepam on r.m.s. noise (Fig. 4.5 A),
further suggests that the receptors passing this tonic current are devoid of the γ
subunit.
There were no significant variations of the r.m.s. current noise recorded from
CA1 PCs (Kruskal Wallis test, p=0.138, 8 d.f) or NAcc MSNs (Kruskal Wallis
test p=0.367, 8 d.f.) across any of the treatment groups (Table 4.6). There were
no trends for any effect of THDOC or diazepam on r.m.s. noise recorded from
5-6 NAcc MSNs of any genotype, so analyses were not extended to the
remaining 44 recordings.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 125
Figure 4.5 – Homozygous mutation α2Q241M specifically disrupts neurosteroid
potentiation of tonic currents in DG GCs
A. Bar chart detailing the change in r.m.s. current noise (mean ± s.e.m.) after
equilibration of DG GCs with THDOC or diazepam. ** - p<0.01 for effect of THDOC
vs. mock, ++ - p<0.01 for diazepam vs. THDOC, ## - p<0.01 for wt vs. hom (ANOVA,
multiple comparisons corrected with Benjamini-Hochberg adjustment).
B. Representative examples of r.m.s. current noise after equilibration under the
defined condition for wt and hom DG GCs.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 126
Tonic currents: drug effects wt het hom
CA1
PCs
mock -0.23 ± 0.13 (n=6) -0.18 ± 0.07 (n=4) -0.05 ± 0.15 (n=3)
50 nM THDOC -0.16 ± 0.09 (n=4) -0.14 ± 0.11 (n=5) 0.15 ± 0.17 (n=5)
100 nM THDOC 0.10 ± 0.09 (n=10) 0.69 ± 0.37 (n=5) 0.31 ± 0.20 (n=5)
DG GCs
mock 0.15 ± 0.05 (n=5) - 0.12 ± 0.08 (n=7)
100 nM THDOC 0.70 ± 0.14 (n=7)** - 0.18 ± 0.10 (n=7)##
500 nM diazepam 0.14 ± 0.06 (n=7)++
- 0.18 ± 0.10 (n=7)
NAcc
MSNs
mock -0.26 ± 0.10 (n=6) -0.08 ± 0.03 (n=6) 0.13 ± 0.27 (n=6)
100 nM THDOC -0.06 ± 0.18 (n=6) 0.15 ± 0.12 (n=6) -0.10 ± 0.11 (n=6)
500 nM diazepam -0.06 ± 0.07 (n=6) -0.07 ± 0.14 (n=6) 0.08 ± 0.04 (n=5)
Table 4.6 – Comparing changes in tonic currents after equilibration with THDOC
or diazepam
Summary of the changes (mean ± s.e.m.) in root mean square current noise (pA) after
equilibration with drug in CA1 PCs, DG GCs or NAcc MSNs. ** - p<0.01 for effect of
THDOC vs. mock, ++ - p<0.01 for diazepam vs. THDOC, ## - p<0.01 for wt vs. hom
(ANOVA, multiple comparisons corrected with Benjamini-Hochberg adjustment).
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 127
4.3. Discussion
4.3.1. Effects of α2Q241M on baseline synaptic and tonic GABA transmission
The α2Q241M mutation has been shown to specifically ablate neurosteroid
potentiation at α2 3γ2S receptors, without effect on responses to GABA
(Section 3.2.1; Hosie et al., 2006; Hosie et al., 2009). Furthermore this mutation
has been successfully introduced into our transgenic mouse strain without any
obvious changes to the expression of GABAA receptor subunits α1-α5 (Sections
3.2.4 and 3.2.5). One may therefore predict that the mutation would have no
effect on baseline inhibitory neurotransmission within the transgenic mouse. In
fact, results are to the contrary: baseline sIPSCs in knock-in mice decay faster
than those of wt littermates, reducing the charge transfer per synaptic event.
This does not necessarily indicate a compensatory alteration in response to the
α2Q241M mutation. Indeed the unchanged IPSC amplitude suggests that a similar
number of GABAA receptors are expressed at the synaptic site, and/or that a
similar amount of GABA is released per synaptic event. Unaltered sIPSC
frequency would also suggest a lack of presynaptic alterations in the knock-in
mice. We propose that the difference in baseline decay times is an indication for
a role of endogenous neurosteroids in modulating inhibitory neurotransmission
within the slice. There is already a precedent for such a role because inhibiting
neurosteroid synthesis reduced the decay times of IPSCs recorded from
pyramidal neurons of the neocortex (Puia et al., 2003). Our results of faster
decay times for sIPSCs from cells of hom α2Q241M mice may therefore identify a
role for α2-type GABAA receptors in responding to endogenous neurosteroids. If
our model is correct, the difference between wt and hom IPSC decays should
be ablated by removing endogenous neurosteroids from the slices.
Interestingly, some effects of the α2Q241M mutation on baseline sIPSC
parameters can also be seen in the heterozygous state (see Section 4.2.1).
There are two α subunits per receptor (Fig. 1.1), and so two neurosteroid
potentiation sites. Assuming that both copies of the subunit, α2M and α2Q, are
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 128
expressed in equal amounts and co-assemble in an unbiased manner within
cells of het animals, we would expect a binomial expression pattern of four
combinations of wild-type and mutant copies of α2 subunits: α2Qα2Q, α2Qα2M,
α2Mα2Q and α2Mα2M. By comparing fits of one- and two-site models to
experimental data, Hosie et al. (2009) suggested that one-site occupation is
sufficient for full neurosteroid potentiation. Furthermore, using concatameric
assemblies to control the GABAA receptor composition, Bracamontes and
Steinbach (2009) showed that both α subunits are capable of contributing to
neurosteroid potentiation. Effects of mutating each site individually were either
small or insignificant, and potentiation by neurosteroid was only completely lost
if both sites were mutated (Bracamontes & Steinbach, 2009). We would
therefore expect het animals to respond ‘normally’ (like wt) to the endogenous
neurosteroids in the slice (since only one in four receptor combinations is
expected to be the insensitive double-mutant, α2Mα2M). However, if there is a
cell-to-cell variation in the relative expression levels of the neurosteroid-
insensitive (α2Mα2M) and sensitive (α2Qα2Q, α2Qα2M, α2Mα2Q) receptor
combinations in heterozygotes, cells with a higher proportion of α2Mα2M
receptors would be expected to have a diminished response to endogenous
neurosteroids within the slice. Our data on subunit expression cannot
distinguish the wt and mutant alleles, but a quantitative PCR-based approach
may help determine the relative expression levels of the two alleles in
heterozygotes.
CA1 PCs, DG GCs and NAcc MSNs all showed very little response to
bicuculline, suggesting that baseline tonic currents are small or absent from
these cells under the conditions of our experiments. This is perhaps not a
surprise: other investigators frequently require bath application of GABA, or a
GABA reuptake inhibitor (such as NO-711), in order to achieve a consistent
tonic current measurement in CA1 PC and DG GCs (e.g. Semyanov et al.,
2003). We have used neither manipulation, in order to preserve slices in a near-
physiological state. Some investigators have observed tonic currents in these
cells without altering extracellular GABA levels (e.g. Bai et al., 2001; Nusser &
Mody, 2002); it is difficult to say what condition may underlie the difference
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 129
between these investigations and ours. Perhaps there are some age- or sex-
related variations in tonic currents; for example the expression of δ-subunit
containing receptors, which dominate in DG GC tonic currents (see Section
1.1.1), increases with increases with age (Laurie et al., 1992b). Our
investigation focussed on young (P18-30) males, and we cannot discount the
presence of more robust tonic currents at older ages, or in females, in these cell
types within our mouse strain. In our investigation, the α2Q241M mutation has no
effect on the magnitude of the baseline tonic current (Table 4.2). This would
imply that, at least under the conditions of our experiment, there is no
involvement of GABAA receptor α2 subunits (or neurosteroid potentiation at
these subunits) in generating baseline tonic currents. Furthermore, unchanged
tonic currents would be consistent with a lack of compensatory changes in
expression of extrasynaptic GABAA receptors in the knock-in mice.
4.3.2. α2Q241M specifically diminishes neurosteroid potentiation of synaptic
transmission
The clearest effect of THDOC and diazepam on the sIPSCs is a prolongation of
decay times, which is seen in recordings from all three cell types. The lack of
effect on IPSC amplitude suggests that GABA release during phasic
transmission is already saturating postsynaptic receptors, and is consistent with
observations of others (e.g. Otis & Mody, 1992). Effects on sIPSC area are
consistent with those on decay time, but these effects are less robust, possibly
because area depends on a combination of decay time (increases in all cells)
and amplitude (although the average suggests no change, within individual
cells, amplitude can increase, decrease or remain unchanged). A lack of effect
of THDOC on inter-event interval is consistent with a lack of presynaptic effects
of THDOC in the hippocampus and NAcc.
The strong effect we observe for 100 nM THDOC on w from wt DG GCs may
appear at odds with reports elsewhere of insensitivity of these cells to THDOC
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 130
(Stell et al., 2003) and their low sensitivity to allopregnanolone (Belelli & Herd,
2003; Harney et al., 2003). However, the neurosteroid sensitivity of IPSC decay
time varies with age, receptor subunit composition, cellular differences in
metabolism and relative kinase/phosphatase activities (see Section 1.2.3). Any
one of these factors may account for the discrepancies between these
observations. For example, DG GC IPSC sensitivity to neurosteroid declines
with age in rats (Cooper et al., 1999; Mtchedlishvili et al., 2003), being highest
during the first two postnatal weeks; responses to ‘physiological’ levels of
THDOC (50 nM and 100 nM) are lost soon after (P17-21 (Cooper et al., 1999)).
Perhaps neurosteroid sensitivity declines at around P30 in mice, which would
explain why IPSCs are responsive in our work (P18-30 mice) but not in that of
Stell et al. (2003) (P30-P181 mice).
Since the α2Q241M mutation ablates neurosteroid potentiation of α2β3γ2S
receptor function in HEK293 cells, one would predict potentiation of GABAergic
sIPSCs to also be defective in α2Q241M knock-in mice. Focussing on the sIPSC
w response to THDOC, data are consistent with this prediction: sIPSC decay
prolongation by 100 nM THDOC is diminished in hom knock-ins all three cell
types examined. In NAcc MSNs and DG GCs, the response is ablated (i.e. no
different to the responses of ‘mock’ treated cells); in CA1 PCs, neurosteroid
sensitivity is significantly diminished, but not completely lost. The most likely
explanation for residual response in CA1 PCs is the retained neurosteroid-
potentiation functionality of other α subunit isoforms in these cells (α1, α3, α4
and α5 (Fig 3.4A; Wisden et al., 1992; Prenosil et al., 2006)). It would therefore
seem likely that α2-type GABAA receptors are responsible for approximately
50% of sIPSCs recorded from CA1 PCs (decay time prolongation by 100 nM
THDOC: 16% in hom vs. 30% in wt), and for the majority of sIPSCs in DG GCs
and NAcc MSNs. Whilst the differential response in cells from hom mice could
alternatively represent a compensatory change in response to the α2Q241M
mutation, subunit expression data would argue against this notion (Section
3.3.2). Furthermore, the IPSC decay time prolongation by 500 nM diazepam is
undiminished by the α2Q241M mutation in DG GCs and NAcc MSNs (Fig. 4.4).
These data are consistent with unaltered function of synaptic GABAA receptors
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 131
(i.e. support a lack of compensatory changes). We therefore propose a key role
for α2-type GABAA receptors in overall neurosteroid response of sIPSCs.
Neurosteroid and diazepam modulation of sIPSCs recorded from cells of het
animals are generally the same as those observed in cells from wt mice. The
response of het NAcc MSNs to 100 nM THDOC, assessed by w prolongation or
by increased event area, is no different to that seen in wt littermates. Similarly,
het NAcc MSNs show a wt-like response to 500 nM diazepam. The response of
het CA1 PCs to 100 nM THDOC are the same as for wt (although pairwise
comparisons find no significant difference for het vs. hom and for het vs. wt).
These data appear to be at odds with the effects of het α2Q241M on baseline
IPSC characteristics discussed in Section 4.3.1; the reason for this discrepancy
remains unclear.
Interestingly, in NAcc MSNs from heterozygotes, the response to 500 nM
diazepam is significantly higher than that for 100 nM THDOC. It is not clear why
this is the case, because these doses of modulator achieve the same degree of
potentiation of wt α2-type GABAA receptors expressed in HEK293 cells (Fig. 3.2
A). However, the diazepam and THDOC sensitivity of heterozygous α2Q241M
mutant receptors has not been examined in HEK293 cells; perhaps the
concatamer-based approach of Bracamontes and Steinbach (2009) should be
utilised to examine the effects of mutating a specific single neurosteroid
potentiation site within the pentamer on diazepam sensitivity. Slice
electrophysiology could also be extended to determine if this effect is specific to
NAcc MSNs: the diazepam sensitivity of CA1 PCs of all three genotypes should
be examined, whilst heterozygous animals should be added to the work in DG
GCs.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 132
4.3.3. α2Q241M reveals a role for α2-type GABAA receptors in tonic currents
Application of 100 nM THDOC significantly increased the r.m.s. current noise
recorded from DG GCs of wt mice (Fig. 4.5, Table 4.6). This increase in noise is
of GABAA receptor origin, because it is reversed by application of 20 µM
bicuculline (Fig. 4.5 B). This effect of THDOC is not apparent in DG GCs in
slices from hom animals, suggesting that it involves neurosteroid potentiation at
α2-type receptors. The lack of effect of 500 nM diazepam on r.m.s. noise (Fig.
4.5 A), further suggests that the receptors passing this tonic current are devoid
of the γ subunit (Pritchett et al., 1989). Potential receptor combinations
accounting for this tonic current could therefore be α2βn and α2βnδ.
There were no significant effects of either THDOC or diazepam on the r.m.s.
current noise recorded from CA1 PCs or NAcc MSNs. Under the conditions of
our experiments, therefore, it seems that tonic GABA currents in both cell types
are carried by receptors lacking the γ subunit (which would bestow diazepam
insensitivity). It has been suggested that neurosteroid insensitivity of tonic
currents indicates that receptors lack the δ subunit (Stell et al., 2003), but 100
nM THDOC clearly has an effect on the synaptic (probably γ-type) receptors in
the same cells, so even a γ-subunit-mediated tonic current would be expected
to respond to THDOC. Perhaps the neurosteroid insensitivity of the
extrasynaptic receptors in these cell types can be attributed to their
phosphorylation state under the conditions of our experiment.
4.3.4. Limitations of the electrophysiological characterisation
Any experimental approach to measuring neurotransmission will itself perturb
normal function, and the further removed the approach is from the intact brain,
the greater the potential for departure from normality. Work in this thesis utilises
brain slice tissue, which represents a state closer to in situ physiology than
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 133
using dissociated neurons in culture, because cells are maintained in their
native local milieu, and many of their original synaptic inputs are retained. The
properties of synaptic and tonic currents recorded form cells in brain slices can
be influenced by the recording temperature, holding voltage, Cl- loading and the
addition of compounds that influence the ambient GABA concentration (Otis &
Mody, 1992; Cooper et al., 1999; Overstreet et al., 2000; Semyanov et al.,
2003; Semyanov et al., 2004; Houston et al., 2009a). Importantly, manipulating
extracellular GABA levels can determine which subunits contribute to that
current, with a more prominent role for α5-type receptors in CA1 PCs at
elevated GABA concentrations (Scimemi et al., 2005; Prenosil et al., 2006). It is
therefore possible that the roles we propose for α2-type GABAA receptors only
hold under the particular experimental conditions used. It would be interesting to
expand our study of tonic current to conditions that will raise ambient GABA
levels, to see if the neurosteroid-stimulated α2-dependent tonic current is
retained. It would also be interesting to extend the investigation, to screen for
the effects of α2Q241M within different cell types and at different developmental
stages, particularly where GABAergic currents may be excitatory (Ben-Ari,
2002; Szabadics et al., 2006; Chiang et al., 2012).
Most experiments have utilised 100 nM THDOC to potentiate GABAergic
transmission within the slice. This concentration represents the higher end of
that which has been measured physiologically (Paul & Purdy, 1992; Concas et
al., 1998), and some investigators suggest it is better to focus on lower
concentrations (less than 50 nM) to represent those seen more commonly
during physiology (Harney et al., 2003; Stell et al., 2003). Interestingly, in our
work, the 50 nM response of hom CA1 PCs does not seem to be lower than that
in recordings from het and wt CA1 PCs (Fig. 4.3), which may suggest that the
α2 subunit is only important in responses to higher neurosteroid levels.
However, it is difficult to know whether the whole-tissue estimates of
neurosteroid levels directly correlate with those found at the level of individual
cells. Endogenous neurosteroids act in a paracrine manner, and applied
neurosteroids can accumulate within cells (Li et al., 2007), such that the
concentration of compound in the cell surface membrane (i.e. at the level of the
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 134
GABAA receptor) may be in excess of that applied in the bath (Chisari et al.,
2010). Furthermore, Belelli and Herd (2003) find that local degradation to
inactive metabolites may shape the response of different cells within
hippocampal slices to the same concentration of bath-applied neurosteroid.
Whether our observed neurosteroid-response deficit relates to physiological
levels of these compounds therefore awaits methods to accurately measure
their concentrations at synapses. Encouragingly, however, there are clear
baseline effects of α2Q241M on IPSC decay time. We propose that this identifies
an endogenous neurosteroid tone within CA1 PCs, DG GCs and NAcc MSNs,
that influences synaptic neurotransmission by function at α2-type GABAA
receptors.
4.4. Conclusions
1. Baseline sIPSCs are faster-decaying in several cell types from hom mice
vs. wt littermates, suggesting a role for endogenous neurosteroids in
setting the duration of inhibitory synaptic transmission. By extension,
hom mice are predicted to have a diminished inhibitory tone in vivo.
2. IPSC responses to diazepam are unperturbed by the α2Q241M mutation in
brain slices from the knock-in mice, consistent with a lack of
compensatory alterations in the mouse strain.
3. Functional potentiation sites on GABAA receptor α2 subunits are required
for a full response of synaptic events to neurosteroids in CA1 PCs, DG
GCs and NAcc MSNs.
Chapter 4: Electrophysiological consequences of losing neurosteroid potentiation at GABAA receptor α2 subunits 135
4. In DG GCs, GABAA receptor α2 subunits appear not to be involved in
baseline tonic currents, but may be recruited to pass such currents at
high THDOC concentrations. The receptors involved are likely to be peri-
or extra-synaptic γ-subunit-lacking receptor combinations (α2βn and
α2βnδ).
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 136
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M
knock-ins
5.1. Introduction
5.1.1. Behaviour of α2Q241M mice defines the physiological and pharmacological
roles of neurosteroids acting at GABAA receptor α2 subunits
The α2Q241M mice generated in this study provide a unique opportunity to
elucidate the roles of α2-type GABAA receptors in mediating responses to both
endogenous and injected neurosteroids. Levels of endogenous neurosteroids
fluctuate during a number of normal and pathological settings, being raised by
stress (Purdy et al., 1991; Barbaccia et al., 1996) and diminished in depression
(Romeo et al., 1998; Uzunova et al., 1998; Strohle et al., 1999; Strohle et al.,
2000), for example. Differences in baseline behavioural phenotypes of the
α2Q241M knock-ins, compared to wild-type littermates, are likely to be a
consequence of losing regulation by these endogenous neurosteroids at α2-
type GABAA receptors, especially given that we have not found any
compensatory changes in expression of GABAA receptors α1-α5 that could
account for changes in behaviour (Chapter 3).
Injected neurosteroids have a number of effects, including anxiolysis (Crawley
et al., 1986; Wieland et al., 1991), antidepression (Khisti et al., 2000), analgesia
(Winter et al., 2003), and sedation (Mendelson et al., 1987), all of which imply
that neurosteroids could prove useful therapeutics for a number of nervous
system disorders. It remains to be determined whether, as for benzodiazepines,
different GABAA receptor α-subunit isoforms are responsible for each specific
behavioural effect of neurosteroids. Examining responses to injected
neurosteroids in α2Q241M mutant mice will demonstrate which behavioural
effects listed above depend on potentiation at the GABAA receptor α2 subunit.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 137
Behavioural studies in this thesis have focussed on paradigms that assess
anxiety-like and depression-like behaviours within the α2Q241M knock-in mouse
strain.
5.1.2. GABAA receptor α2 subunits and neurosteroids in anxiety
GABAA receptor α2 subunits are key components of the anxiety circuitry in the
CNS (see Section 1.3.2). Fundamental observations are that α2 expression is
strong in brain areas linked to mood and anxiety, including cortex, hippocampus
and amygdala (Wisden et al., 1992; Sperk et al., 1997; Pirker et al., 2000;
Sieghart & Sperk, 2002; Mohler, 2006a), and that knock-out of this subunit
causes an anxiety phenotype (Dixon et al., 2008). Furthermore, the anxiolytic
response to benzodiazepines is thought to be mediated by α2-type GABAA
receptors (Low et al., 2000). If GABAA receptor α2 subunits are also central to
the anxiolytic response to endogenous neurosteroids, one would predict a
baseline anxiety phenotype for hom α2Q241M mutants, which will be incapable of
responding to these compounds at α2 subunits. Moreover, this defect would be
predicted to impair anxiolytic responses to injected neurosteroids. These
predictions were tested using two behavioural screens for anxiety-like
behaviour: the elevated plus maze and light-dark box tests.
The elevated plus maze has been validated over many years as a screen for
anxiety-like behaviour in mice (Lister, 1987; Hogg, 1996). This test, described in
Section 2.7.2, exploits the aversive properties of open spaces. Mice are faced
with a choice between exploring regions open to the environment and ‘safer’
enclosed regions, which supply thigmotactic cues. Drugs known to be anxiolytic
in humans, including benzodiazepines, increase the number of entries into and
percentage time spent on the aversive open arms (Pellow et al., 1985).
Neurosteroids, such as allopregnanolone and THDOC, have equivalent effects
on plus maze behaviour (Crawley et al., 1986; Rodgers & Johnson, 1998).
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 138
The light-dark box also comprises an environment divided into aversive (light
zone) and safe (dark zone) areas. As with the elevated plus maze, the test was
validated using benzodiazepine anxiolytics (Crawley & Goodwin, 1980;
Blumstein & Crawley, 1983), which release innate inhibition on mouse
exploratory behaviour, and so increase exploration of both zones, especially the
aversive light zone. Different investigators find divergent effects of anxiolytics on
the various parameters scored within this paradigm, leading to debate over
which parameters are the best measures of anxiety (e.g. see discussion by
Hascoet & Bourin, 1998). Anxiolytic responses can include some or all of the
following changes: increased time to first exit the light zone, increased time
spent in the light zone, and an increase in exploratory locomotion within both
zones. All of these parameters are therefore considered in this study.
5.1.3. GABAA receptor α2 subunits and neurosteroids in depression
Common mechanisms may underlie anxiety and depression, because these
disorders are frequently co-morbid (Hirschfeld, 2001; Nutt et al., 2006). One
proposal is that deficiencies in GABAergic transmission could account for both
disorders (see Section 1.4.1). Of particular note for this study, are observations
that α2-/- mice have phenotypes consistent with both anxiety and depression
(Dixon et al., 2008; Vollenweider et al., 2011) and that several limbic regions
whose structure and/or activity is altered in depression, including the
hippocampus, amygdala and basal ganglia (Sheline, 2003; McCabe et al.,
2009), are areas rich in GABAA receptor α2 subunit expression (Wisden et al.,
1992; Sperk et al., 1997; Pirker et al., 2000; Sieghart & Sperk, 2002; Mohler,
2006a). These data would support a role for α2-type GABAA receptors in
depression.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 139
Neurosteroids have also been linked with depression (see Section 1.4.2). The
two key points being a negative correlation between endogenous
allopregnanolone levels and depression symptoms in humans (Romeo et al.,
1998; Uzunova et al., 1998; Strohle et al., 1999; Strohle et al., 2000), and that
the antidepressant function of injected neurosteroids in animal models involves
GABAergic signalling (Khisti et al., 2000). We therefore propose that insufficient
neurosteroid potentiation at GABAA receptor α2 subunits could underlie co-
morbid anxiety and depression. This postulation was tested here by measuring
depression-related behaviour in the α2Q241M knock-in mouse strain. As with
anxiety phenotyping, experiments examined baseline behaviour of wt and hom
α2Q241M mice, to expose any role for endogenous neurosteroids at α2-type
GABAA receptors. Responses to injected THDOC were also measured, to
elucidate any therapeutic potential for neurosteroids at these receptors.
In this study, depression-related phenotypes were screened using the tail
suspension test paradigm. Mice are held upside-down by their tail for six
minutes, during which time they fluctuate between two types of behaviour:
periods of motion that reflect attempts to escape, and periods of immobility
thought to represent ‘behavioural despair’. This approach was developed in the
1980s (Steru et al., 1985; Thierry et al., 1986), and was validated by showing
that known antidepressant drugs decreased behavioural despair (i.e. reduced
‘immobility time’) in mice. Increases in immobility time would therefore represent
phenotypic depression. The tail suspension test was chosen in preference to
the related forced swim test (Porsolt et al., 1978) because it appears to be more
sensitive to low doses of antidepressant, and does not induce hypothermic
stress (see discussion in Thierry et al., 1986).
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 140
5.2. Results
5.2.1. Endogenous neurosteroids act via GABAA receptor α2 subunits to
determine basal anxiety levels
In both the elevated plus maze (Fig. 5.1) and light-dark box (Fig. 5.2) tests, hom
knock-in mice have an anxiety-like phenotype under basal conditions. Anxiety
inhibits the tendency of mice to explore novel environments, and is particularly
manifest in reduced exploration of the aversive portions of the equipment used
in these tests.
In the elevated plus maze hom animals spend significantly less time on the
open arms than wt animals (Fig. 5.1 A), and tend to make fewer entries onto
these arms (Fig. 5.1 B), although this latter trend does not reach significance (p
= 0.17, paired t-test before correction for multiple comparison). Importantly,
these effects are not explained by differences in activity within the maze: closed
arm entries (Fig. 5.1 C) and total arm entries (Fig. 5.1 D) are invariant across
genotypes. Heterozygous animals were included in this paradigm, and
interestingly appear to have an anxiety phenotype intermediate between wt and
hom animals, with data being neither significantly different to wt nor to hom
mice.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 141
Figure 5.1 – α2Q241M confers an anxious phenotype in the elevated plus maze
A. and B. Bar charts detailing the percentage time spent in the open arms (A) and the
percentage of entries onto open arm (B) for each genotype (n=9 animals per
genotype). There is a tendency for hom and het animals to spend less time on, and
make fewer entries onto the open arms compared to wt. Hom animals spend
significantly less time on open arms than wt littermates (* - p<0.05 paired t-test,
Bonferroni corrected for multiple comparisons).
C. and D. Activity measures – total arm entries (C) and closed arm entries (D) – are
unchanged across genotypes.
Elevated plus maze findings are corroborated by the behaviour of uninjected
mice in the light-dark box paradigm. This test focussed on comparing wt and
hom animals only. Several parameters scored in this procedure are consistent
with hom mice being anxious relative to wt littermates. Firstly, hom mice leave
the light zone more quickly at the start of the test than wt mice (Fig. 5.2 A). As
well as this active avoidance response, anxiety in hom mice also increased their
passive avoidance responses, manifesting in a decreased number of returns to
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 142
the light zone (i.e. fewer transitions between zones – Fig. 5.2 B) and a reduced
time spent in the light zone (and correspondingly more time in the dark zone –
Fig. 5.2 C).
Figure 5.2 – α2Q241M confers an anxious phenotype in the light-dark box
Homozygotes are faster to exit the light zone (A), make fewer transitions between the
two zones (B), spend more time in the dark zone (C) and tend to show less exploratory
activity (D) within the box than wt littermates (n=8 animals per genotype). Significant
differences between genotypes are marked: * - p<0.05 and ** - p<0.01 (unpaired t-
test).
The reduced exploratory activity of hom mice within the dark zone is also
consistent with an anxiety phenotype (Fig. 5.2 D). Although the exploratory
activity in the light zone is no different between wt and hom animals (Fig. 5.2 D),
all of the other parameters are in line with an anxiety phenotype in hom animals.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 143
Data from both tests therefore support the prediction that endogenous
neurosteroids modulate baseline anxiety, at least in part, by potentiation at α2-
type GABAA receptors.
5.2.2. GABAA receptor α2 subunits mediate the anxiolysis of injected
neurosteroids at low doses
Having demonstrated a role for GABAA receptor α2 subunits in anxiolytic
responses to endogenous neurosteroids, the investigation was extended to ask
whether this subunit is also involved in mediating the anxiolytic response to
exogenously applied neurosteroids. Plus maze and light-dark box tests were
repeated, using mice pre-injected with the neurosteroid THDOC, at doses
previously known to be anxiolytic in both tests (Wieland et al., 1991; Rodgers &
Johnson, 1998).
In the elevated plus maze, both doses of THDOC (10 mg/kg and 20 mg/kg) are
anxiolytic in wt animals, increasing percentage time on open arms (Fig. 5.3 A)
and percentage open arm entries (Fig. 5.3 B) relative to vehicle injected mice.
The total number of arm entries increases at 20 mg/kg THDOC (Fig. 5.3 D), but
there is no change in closed arm entries (Fig. 5.3 C). Although the former could
suggest an increased locomotor activity, some investigators have suggested
closed arm entries is a better activity measure, because it is independent of the
(anxiety-related) number of open arm entries (Rodgers & Johnson, 1995).
THDOC injection is also anxiolytic in hom mice on the elevated plus maze, but
only at the higher (20 mg/kg) dose: at 10 mg/kg THDOC, hom animals spend
significantly less time on and make significantly fewer entries into open arms
than wt littermates (Fig. 5.3 A, B), and these parameters are no different to
those for vehicle-injected homozygotes. Such a rightward shift in the anxiolytic
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 144
dose-response curve for α2Q241M mice is consistent with our prediction of
reduced anxiolytic potency of neurosteroids.
Figure 5.3 – α2Q241M influences behavioural responses to injected THDOC on the
elevated plus maze
A. and B. Bar charts detailing the effect of THDOC injection on percentage time spent
on the open arms (A) and the percentage open arm entries (B) for each genotype.
C. and D. Activity measures – closed arm entries (C) and total arm entries (D) – are
potently reduced by THDOC in hom animals only.
Data represent the mean (error bars, s.e.m.) of 6-7 animals per group. * - p<0.05, ** -
p<0.01 and *** - p<0.001 for effect of THDOC vs. vehicle (veh); # - p<0.05, ## - p<0.01
and ### - p<0.001 for effect of genotype at a given dose of THDOC. Comparisons
were either ANOVA, with multiple comparisons corrected with Benjamini-Hochberg
adjustment (B, C), or all pairwise Behrens-Fisher comparisons (A, D).
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 145
Unexpectedly, hom animals respond to THDOC with profound reductions in
locomotion: both doses significantly decreased closed arm entries (Fig. 5.3 C)
and total arm entries (Fig. 5.3 D) when compared to vehicle treated animals, or
to wt animals at equivalent THDOC dose. Whilst these activity effects could cast
doubt over the genotypic difference in anxiety measures, the percentage open
arm entries measure inherently accounts for total arm entries made during the
test (i.e. the reduced activity is accounted for in Fig. 5.3 B). The conclusion that
anxiolytic potency of neurosteroids is diminished in α2Q241M mice therefore still
stands. However, THDOC’s ataxic effects pose some serious problems with the
light-dark box approach, as will be discussed below. The mechanism underlying
these activity effects is explored in Section 5.2.5.
In the light-dark box procedure, as with the elevated plus maze, injections of
THDOC have classical anxiolytic effects in wt animals. The 20 mg/kg dose
increases time to first exit the light zone compared to vehicle-treated animals
(Fig. 5.4 A), consistent with an abated active avoidance response to the
aversive light zone. This dose of THDOC also releases inhibition on exploratory
activity in the wt mice, leading to increased transitions between the two zones
(Fig. 5.4 B) and increased line crossings within each zone (Fig. 5.4 D).
However, THDOC’s anxiolytic effects are not manifest in an increased time
spent in the light zone (Fig. 5.4 C).
Surprisingly, THDOC injection into hom mice has apparent anxiogenic effects in
the light dark box, including significant reductions in transitions between zones
(Fig. 5.4 B) and time spent in the light zone (Fig. 5.4 C). THDOC also tends to
reduce exploratory activity in both zones; although the effect only reaches
significance in the dark zone at 20 mg/kg vs. vehicle, activity is significantly
lower than for wt animals at both doses of THDOC (Fig. 5.4 D). Given the
potent ataxic effects of THDOC in hom animals within the elevated plus maze,
however, it is difficult to determine whether these are truly anxiogenic effects, or
an artefact of altered motor activity. For example, decreased time in the light
zone may represent mice sleeping/resting in the dark zone, rather than avoiding
the light zone.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 146
Figure 5.4 – α2Q241M alters response to THDOC injection in the light-dark box
Bar charts detailing the effect of intraperitoneally-injected THDOC on time to exit the
light zone (A), the number of transitions between the two zones (B), time spent in the
light zone (C) and the exploratory activity within each zone (D).
Data represent the mean (error bars, s.e.m.) of 6-9 animals per group. * - p<0.05, ** -
p<0.01 and *** - p<0.001 for effects of THDOC vs. vehicle (veh); + - p<0.05 for effect of
20 mg/kg vs. 10 mg/kg THDOC; # - p<0.05, ## - p<0.01 and ### - p<0.001 for effect of
genotype at a given dose of THDOC. Comparisons were by ANOVA, with multiple
comparisons corrected with Benjamini-Hochberg adjustment, except dark zone
exploratory activity, which was subject to all pairwise Behrens-Fisher comparisons.
Unlike in the elevated plus maze, there is no easy means to correct for
THDOC’s activity effects in the light-dark box. Interestingly, THDOC-treated
hom mice still exit the light zone very quickly (Fig. 5.4 A). Apparently, the initial
need to escape from the aversive light zone is sufficient motivation for active
avoidance, despite the drug’s activity-reducing effect. Time to exit the light zone
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 147
at 20 mg/kg THDOC is no different to that for vehicle treated mice – consistent
with a lack of anxiolytic effect of the drug in hom animals. Therefore, light-dark
box results may well concur with those from the elevated plus maze.
Scoring for the light-dark box test was performed in two minute bins, allowing
any time-dependent effects to be probed. Habituation of mice to the new
environment of the box manifests in a tendency for reduced exploratory activity,
in either zone, with increasing time. Habituation of wt and hom animals occurs
at similar rates following vehicle injection (Fig. 5.5 A), but the effects of THDOC
treatment on activity timecourses diverge between genotypes. THDOC
counteracts habituation in wt mice (Fig. 5.5 B), and at the higher dose actually
increases activity over time (compare bins 2/3 with bin 1). Conversely, THDOC
accelerates habituation of hom mice (Fig. 5.5 C), with significant activity
reductions occuring at earlier time bins than vehicle-treated controls.
Interestingly, THDOC does not significantly affect activity of either genotype in
the first scoring bin (0-2 min timepoint of test, effects of dose: wt, p = 0.20; hom,
p = 0.22). The time spent in each zone during this time bin may therefore be
used as an anxiety measure without confound from activity effects (Fig. 5.5 D).
A trend consistent with THDOC-mediated anxiolysis (increased time in light
zone / decreased time in dark zone) is seen in wt animals, although this does
not quite reach significance. Conversely, THDOC appears anxiogenic in hom
animals (less time in light zone / more time in dark zone) with significant effects
of 20 mg/kg THDOC vs. vehicle-treated homs and vs. 20 mg/kg THDOC-treated
wt mice. When the same analysis is applied to data from uninjected animals
(Fig. 5.5 E), hom mice still tend toward an anxious phenotype, although the
difference between genotypes is not quite significant (p = 0.09, t-test). Data in
Fig. 5.5 D therefore seem consistent with elevated plus maze findings – the
α2Q241M mutation has disrupted the anxiolytic response to injected neurosteroid.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 148
Figure 5.5 – Timecourse analysis of light-dark box results
A – C. Time courses for activity (grid lines crossed) in both zones of the light-dark box for
mice injected with 2-hydroxypropyl-β-cyclodextrin vehicle (A) show a trend to decrease.
This decrease is prevented by THDOC in wt animals (B), and accelerated by THDOC in
hom animals (C). Activity within the first two minutes of the test (bin 1) is not significantly
affected by THDOC in either genotype. Presented data represent the mean (error bars,
s.e.m.) of 6-9 animals per group. * - p<0.05 and ** - p<0.01 (repeated measures ANOVA
with Dunnett comparisons to bin 1).
D – F. Bar charts detail the time spent in each zone of the light-dark box in the first two
minutes of the test for experiments with THDOC injected (D), untreated (E) and diazepam
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 149
injected (F) mice. Presented data represent the mean (error bars, s.e.m.) of 6-9 animals
per group. *** - p<0.001 for effect of THDOC vs. vehicle (veh); + - p<0.05 for effect of 4
mg/kg diazepam vs. 0.5 mg/kg and 1 mg/kg doses of diazepam; ### - p<0.001 for effect of
genotype at 20 mg/kg THDOC. Comparisons were either by ANOVA, with multiple
comparisons corrected with Benjamini-Hochberg adjustment (F) or non-parametric all-
pairwise Behrens-Fisher comparisons (D).
5.2.3. Anxiolytic effects of pentobarbital and diazepam are retained in α2Q241M
knock-in mice
Results presented above indicate a disrupted anxiolytic response to
neurosteroids in α2Q241M mice. This could be a specific consequence of losing
neurosteroid potentiation at α2-type GABAA receptors, or due to some general
defect in their anti-anxiety circuitry. Sensitivity to pentobarbital and diazepam
potentiation is normal for α2Q241Mβ3γ2S receptors expressed in HEK293 cells,
and for sIPSCs recorded from brain slices of hippocampus and nucleus
accumbens (Fig. 3.1, Fig. 4.4; Hosie et al., 2006). A normal anxiolytic response
to diazepam or pentobarbital would therefore be expected for α2Q241M knock-in
mice. However, refer back to Section 3.3.3 for a discussion of how endogenous
neurosteroids may complicate behavioural approaches. In spite of these
complications, the anxiolytic potencies of pentobarbital and diazepam have
been assessed using the plus maze and light-dark box procedures.
Pentobarbital was more potent in our mouse strain than reported elsewhere
(Lister, 1987): in preliminary tests, 20 mg/kg pentobarbital induced loss of
righting reflex in our strain. Lower intraperitoneal doses were therefore
employed in this test (10 and 15 mg/kg). Although there is a significant increase
in time spent on open arms for wt animals at 15 mg/kg (Fig. 5.6 A), this may
well be a consequence of hyperactivity at this dose (increased total arm entries
– Fig. 5.6 D). Using percentage open arm entries (Fig. 5.6 B) as an anxiety
measure that is independent of altered activity, we see a trend to an increase
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 150
(i.e. anxiolysis) in both genotypes; however two-way ANOVA suggests the
effect of drug is not quite significant (p = 0.06, F = 3.04, 2 d.f.). Encouragingly,
we can be confident that there is no effect of genotype (ANOVA p = 0.32, F =
1.02, 1 d.f.), nor an interaction effect (ANOVA p = 0.64, F = 0.45, 2 d.f.). There
is no effect of drug or genotype on activity as assessed by closed arm entries
(Fig. 5.6 C), although there is a tendency towards increased activity at the
higher dose of pentobarbital.
Because pentobarbital also appeared to be strongly motor impairing at the 15
mg/kg anxiolytic dose – mice showed an uncoordinated gait, or even fell off the
open arms of the maze – this drug was not taken forward to light-dark box
comparisons. Overall, elevated plus maze results suggest that pentobarbital
tends to be anxiolytic with similar potency in both genotypes.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 151
Figure 5.6 – α2Q241M does not influence pentobarbital’s effect on elevated plus
maze behaviour
A. and B. Bar charts detailing the effect of pentobarbital injection on percentage time
spent in the open arms (A) and the percentage open arm entries (B) for each genotype.
C. and D. Activity measures – closed arm entries (C) and total arm entries (D) – reveal
pentobarbital-induced hyperactivity in wt animals only.
Data represent means (error bars, s.e.m.) for 6 animals per group. *** - p<0.001 for
effects of drug vs. saline (all pairwise Behrens-Fisher comparisons).
Diazepam’s anxiolytic effects in the elevated plus maze were seen in both wt
and hom mice, which tend to spend more time on open arms with diazepam
treatment (Fig. 5.7 A). For percentage time on open arms, two-way ANOVA
finds an effect of drug (p = 0.02, F = 4.37, 2 d.f.) but no effect of genotype (p =
0.61, F = 0.27, 1 d.f.) and no interaction effect (p = 1.0, F = 0, 2 d.f.); none of
the individual pairwise comparisons reach significance. Anxiolytic effects of
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 152
diazepam are additionally seen as increased percentage entries onto open
arms (overall Kruskal-Wallis test, p = 0.03), individual pairwise comparisons in
this case also reveal significant effects of drug vs. vehicle (and no differences
between genotypes at any dose – see Fig. 5.7 B). There are no statistically
significant effects of diazepam or genotype on activity measures within the plus
maze (Fig. 5.7 C, D). Results are therefore consistent with a retained anxiolytic
effect of diazepam in α2Q241M mice.
Figure 5.7 – α2Q241M does not influence diazepam’s effect on elevated plus maze
behaviour
A. and B. Bar charts detailing the effect of diazepam injection on percentage time
spent in the open arms (A) and the percentage open arm entries (B) for each genotype.
C. and D. Activity measures – closed arm entries (C) and total arm entries (D) – are
not significantly affected by genotype or diazepam administration.
Data are the mean (error bars, s.e.m.) of 6-9 animals per group. ** - p<0.01 for effects
of diazepam vs. vehicle (veh) (all pairwise Behrens-Fisher comparisons).
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 153
A wider range of diazepam doses was employed in the light-dark box procedure
to probe dose-responses in finer detail. Diazepam’s effects in wt animals closely
mirror the pattern of effects seen by Hascoet and Bourin (1998). We see a
biphasic effect of diazepam on activity in light and dark zones (Fig. 5.8 D): low
doses, 0.5 and 1 mg/kg, tend to increase activity (anxiolytic effect), whilst higher
doses, 2 and 4 mg/kg, tend to decrease activity (sedative effect). Although
increased time to exit the light zone at 2 and 4 mg/kg (Fig. 5.8 A) could
represent anxiolytic effects of the drug, decreased exploratory activity at these
doses implicates a contribution of sedation. Transitions between zones (Fig. 5.8
B) and time spent in the light zone (Fig. 5.8 C) fail to demonstrate diazepam
anxiolysis. The former parameter only revealed sedative effects of higher
diazepam doses, and the latter shows no significant variation with dose.
Homozygous animals appear unresponsive to diazepam in the light-dark box
test, showing no increase in time to exit light zone (Fig. 5.8 A), no change in
transitions between zones (Fig. 5.8 B), no change in time spent in the light zone
(Fig. 5.8 C), nor any alteration in exploratory activity in either zone (Fig. 5.8 D).
Exploratory activity clearly depends on a mixture of anxiolysis (increases
activity) and sedation (decreases activity). The apparent diazepam insensitivity
of exploratory activity in hom animals could result if these mice were more
sensitive to sedation – e.g. sedation at a dose of 1 mg/kg cancelling out the
activity increase expected for anxiolysis. This notion is consistent with the
apparent hom hypersensitivity to sedation with THDOC. However, transitions
between zones and time to exit light zone results are not consistent with ‘hyper-
sedation’ by diazepam, arguing instead that hom α2Q241M mice are simply
insensitive to diazepam in this test. Furthermore, even when focussing only on
the first time bin of scoring (Fig. 5.5 F), an approach that helped to reduce the
problem of THDOC’s activity effects, data still fail to demonstrate an anxiolytic
effect of diazepam.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 154
Figure 5.8 – α2Q241M alters response to diazepam in the light-dark box
Bar charts detailing the effect of intraperitoneally-injected diazepam on time to exit the
light zone (A), time spent in the light zone (B), the number of transitions between the
two zones (C) and the exploratory activity within each zone (D).
Presented data represent the mean (error bars, s.e.m.) of 7-8 animals per group.
Statistically significant differences are highlighted: * - p<0.05 and ** - p<0.01 for effect
of diazepam dose vs. vehicle (veh); + - p<0.05 and ++ - p<0.01 for effect of diazepam
dose vs. 1 mg/kg diazepam; & - p<0.05 and && - p<0.01 for effect of diazepam dose
vs. 0.5 mg/kg diazepam; # - p<0.05 and ## - p<0.01 for effect of genotype at a given
diazepam dose. Comparisons were by ANOVA, multiple comparisons corrected by
Benjamini-Hochberg adjustment.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 155
5.2.4. Knock-in mutation α2Q241M does not confer a depression phenotype
There is no effect of the knock-in mutation on the immobility time of untreated
mice subjected to the tail suspension test (Fig. 5.9 A), which suggests that the
α2Q241M mutation does not induce depression – at least not in the homozygous
state. Injected THDOC has no effect on immobility time of wt mice, and tends to
increase immobility in hom mice at both doses (Fig. 5.9 B). Two-way ANOVA
results indicate an effect of genotype (p = 0.006, F = 8.92, 1 d.f.), no overall
effect of THDOC treatment (p = 0.061, F = 3.08, 2 d.f.) and no interaction effect
(p = 0.316, F = 1.2, 2 d.f.). Unadjusted LSD comparisons suggest significant
effects of 5 mg/kg (p = 0.03) and 10 mg/kg (p=0.03) THDOC vs. vehicle for hom
animals, and a significant difference between immobility time for wt and hom
animals at 10 mg/kg THDOC (p = 0.01); however, none of these values survive
correction for multiple comparisons.
Figure 5.9 – Immobility in the tail suspension test is unaffected by mutation
α2Q241M
A. Baseline behavioural despair (total immobility time) in untreated animals is no
different across genotypes (n=8 animals per genotype).
B. THDOC does not decrease immobility time in wt animals; the trend to increased
immobility time with THDOC dose in hom animals is consistent with ataxic effects
observed in other tests. None of the individual pairwise comparisons reach significance
(n=5-7 animals per group).
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 156
The tendency toward increased immobility in hom animals is consistent with
above observations that THDOC reduces motor activity in these mice, and so is
unlikely to indicate a depressive effect of THDOC injection. The lack of
reduction in immobility in wt mice suggests that THDOC is not antidepressant
within the tail suspension test.
5.2.5. Ataxic effects of THDOC injection in hom mice: not a consequence of
increased motor impairment
Performance of wt and hom mice on an accelerating rotarod was used as a
measure of motor coordination. Scoring assessed the time taken for the mice to
fail the test by either falling off the rod, or passively rotating with it. As can be
seen from Fig. 5.10 A, initial performance on the rod is no different between wt
and hom mice – i.e. the α2Q241M mutation has no inherent motor-impairing
effects.
Mice were next trained to achieve a consistent and reproducible performance
on the rotarod, as detailed in the Section 2.7.5. Performance of these mice was
then assessed following THDOC injection, and expressed for each mouse
relative to his performance on the preceding training day (Fig. 5.10 B). There
was no motor impairing effect of vehicle or THDOC at a dose of 10 mg/kg. An
approximately 20 percent decline in performance is seen for both wt and hom
animals at 20 mg/kg THDOC. Whilst the effect of drug only reaches significance
for hom animals, there is no difference in performance of wt and hom animals at
any dose. In addition, 10 mg/kg THDOC, which has clear activity-reducing
effects in hom animals in other behavioural tests, produces no motor
impairment within the rotarod test. This would suggest that the activity effects of
THDOC cannot be explained by an enhancement of motor impairing effects of
this drug in hom mice.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 157
Figure 5.10 – α2Q241M does not affect performance on the accelerating rotarod
under baseline or THDOC-treated conditions
A. Baseline performance on the rotarod (time to fail the test) is no different across
genotypes.
B. THDOC impairs motor performance on the rotarod (reduces time to failure) only
when administered at a dose of 20mg/kg. This effect reaches significance in hom
animals: * - p<0.05 for effect of 20 mg/kg THDOC vs. vehicle (veh); + - p<0.05 for
effect of 20 mg/kg THDOC vs. 10 mg/kg THDOC; all pairwise Behrens-Fisher
comparisons.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 158
5.3. Discussion
5.3.1. Neurosteroids act via α2-type GABAA receptors to modulate anxiety state
Two behavioural screens for anxiety – the elevated plus maze and the light-dark
box – both demonstrate an anxious phenotype for untreated hom α2Q241M mice
relative to wt littermates. This result is consistent with a model in which
endogenous neurosteroid tone in a naïve untreated mouse regulates its
baseline anxiety state through α2-type GABAA receptors. This conclusion relies
on a lack of compensatory changes that may provide an alternative explanation
for these behavioural differences. In favour of this, no changes in GABAA
receptor protein expression were detected for subunits α1-α5 (see Chapter 3).
Heterozygous animals were included in the elevated plus maze paradigm, and
interestingly appear, by trend, to have an anxiety phenotype intermediate
between wt and hom animals. This appears incongruous with work in brain
slices, where THDOC-mediated prolongation of sIPSC decay time is the same
in cells from het and wt animals (Fig. 4.3, Fig. 4.4). However, het mice have
faster decaying baseline sIPSCs than wt mice (Table 4.1), suggesting the
α2Q241M mutation has some effect when expressed in the heterozygous state.
As described in Section 4.3.1, a single neurosteroid potentiation site is believed
sufficient to achieve the full potentiating effect of neurosteroid, and so only a
quarter of the α2-containing pentamers in het animals should be neurosteroid
insensitive. Perhaps this small deficit is sufficient to produce a phenotypic effect
in some situations. Nevertheless, subsequent experiments focussed on
comparing only wt and hom animals, where we can be sure that neurosteroid
potentiation has been ablated at all α2-type GABAA receptors in the knock-in
mice.
Anxiolytic effects of injected neurosteroids are well established (Crawley et al.,
1986; Wieland et al., 1991), and the stress-induced rise in neurosteroids has
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 159
been purported to be a natural anxiolytic response to restore GABAergic tone
after stress (Purdy et al., 1991; Barbaccia et al., 1998). Here we provide the first
direct evidence for endogenous neurosteroids having an anxiolytic effect in an
unperturbed animal, via α2-type GABAA receptors.
Our suggestion that α2-type GABAA receptors may be mediators of the
anxiolytic response to injected neurosteroid is also supported by the results of
behavioural screening. For the elevated plus maze, a clear rightward-shift in the
dose-response curve for percentage open arm entries indicates a defective
anxiolytic response to THDOC in hom mice. The anxiolytic response to the
higher dose (20 mg/kg THDOC) does not invalidate our model: α2 and α3
isoforms appear to have overlapping roles in benzodiazepine-mediated
anxiolysis and myorelaxation, with α2 mediating low-dose effects, and α3
mediating high-dose effects (see Section 1.3.2). We may be observing a related
phenomenon for THDOC-mediated anxiolysis (i.e. 10 mg/kg acts via α2
subunits, whilst 20 mg/kg can act via α3 subunits). Future experiments to
confirm this hypothesis would be to cross-breed α2Q241M mice with either α3
knock-out mice (e.g. those generated by Yee et al., 2005) or neurosteroid-
insensitive α3Q266M knock-in mice. Mice resulting from such crosses would lose
neurosteroid potentiation at both α2 and α3 isoforms of the GABAA receptor,
and be predicted to demonstrate no anxiolytic response to injected THDOC,
even at high doses.
Modulation of non-GABAergic transmission could also account for the residual
anxiolysis in hom α2Q241M mutants. Whilst GABAA receptors are sensitive to
nanomolar levels of neurosteroids, micromolar levels can also modulate
glutamate (NMDA, AMPA and kainate), glycine, serotonin, nicotinic
acetylcholine, oxytocin, and sigma-type-1 receptors (see review by Rupprecht &
Holsboer, 1999). It is difficult to estimate what brain concentration of THDOC
results from our intraperitoneal injections, which will depend on a number of
factors. Relative absorption and clearance rates will determine the
pharmacokinetic profile, whilst a number of factors – such as propensity to
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 160
cross the blood-brain barrier – will influence the relative distribution in the brain
vs. other tissues. In rats, 30 min following subcutaneous injection of 8 mg/kg
allopregnanolone, brain levels of this compound reach 400 ng per gram tissue
(Vallee et al., 2000). If intraperitoneally injected THDOC were to act similarly in
our mouse strain, and if the amount reaching the brain were directly
proportional to the amount injected, a 20 mg/kg dose of THDOC might be
expected to produce a brain level of approximately 1000 ng/g tissue (i.e. 3
micromoles per kg tissue). Whilst this extrapolation is based on some
oversimplifying assumptions, it would suggest that brain THDOC concentrations
in injected mice could be in the range of the micromolar levels that modulate
non-GABAergic transmission.
THDOCs anxiolytic effects in wt mice subjected to a light-dark box test have
been reported previously (Wieland et al., 1991). We see this effect here as an
increased time to exit the light zone, increased activity in both zones and
increased transitions between zones. The time spent in the light does not
significantly increase, but as alluded to earlier (Section 5.1.2), different
investigators report distinct patterns of drug effects on light-dark box
parameters. For example, no effect of diazepam on this parameter has been
observed by Hascoet and Bourin (1998), depending on mouse strain and
precise conditions of the test. Perhaps under some testing conditions, vehicle-
injected mice are already spending maximal time in the light zone, precluding a
further increase by an anxiolytic.
Drug effects on activity measures complicate matters further. For example, not
all investigators agree with using increased time to exit light zone as an
indication of anxiolysis because it could also indicate sedation. Given that 20
mg/kg THDOC increases wt exploratory activity, the increased time to exit the
light zone probably represents anxiolysis rather than sedation. The increased wt
activity following THDOC injection (increased transitions between zones and
increased exploratory activity) could also represent a stimulant effect of
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 161
THDOC, but the lack of increase in closed arm entries on the elevated plus
maze indicates anxiolysis is the more likely explanation.
THDOC has potent ataxic effects in hom mice, precluding a standard light-dark
box approach to studying anxiety, because it is difficult to disentangle activity
and anxiety effects within this test. There are some indications, however, that
α2Q241M knock-in disrupts anxiolysis. Firstly there is no increase in time to exit
the light zone with THDOC treatment of hom mice (Fig. 5.4 A), whilst activity-
reducing and anxiolytic effects of this drug would both be expected to increase
this duration. It would therefore seem that placement in the light zone is
sufficiently aversive to motivate an active avoidance response in hom mice,
despite the activity-reducing effects of THDOC. Additional support for failure of
anxiolysis in hom mice comes from analysis of the first two minutes of the test,
where activity effects are not significant. Using the relative time spent in light vs.
dark zones as an anxiety measure, wt mice show an anxiolytic-like response to
THDOC, but hom mice demonstrate anxiogenesis. Whilst anxiogenesis was not
seen in the elevated plus maze, these data concur that α2Q241M has disrupted
anxiolysis.
The tests used in this body of work probe only one type of anxiety, relating to
inhibition of exploratory behaviour. This anxiety can also be probed with an
open field apparatus, but one would predict results to be very similar to those of
the light-dark box test – and to suffer the same difficulties when trying to
distinguish emotionality from locomotor/activity effects (Stanford, 2007).
Alternative rodent anxiety screens include conditioned paradigms, where
subjects learn to associate particular non-noxious environmental cues with a co-
presented foot-shock, and subsequently demonstrate fearful responses to these
cues in the absence of foot-shock. There is a precedent for a role of GABAA
receptor α2 subunits in this type of anxiety, because α2-/- mice show an
anxious phenotype in a conditioned emotional response task (Dixon et al.,
2008). It would be interesting to compare α2-/- and hom α2Q241M mice in such a
task, to investigate whether the anxious phenotype of α2-/- mice is a
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 162
consequence of losing neurosteroid potentiation at the α2-type GABAA
receptors.
5.3.2. Anxiolytic effects of pentobarbital and diazepam
To ensure that the disrupted anxiolysis in hom α2Q241M mice is purely a
consequence of lost neurosteroid potentiation at the α2-type GABAA receptors,
and not a general inability of hom mice to respond to any anxiolytic agent,
behavioural studies were extended to include other GABAA receptor
potentiators. If the normal circuitry for anxiety is intact in hom mice, and there
have been no compensatory changes in response to the α2Q241M mutation,
sensitivity to diazepam and pentobarbital should be retained. This is certainly
true for the electrophysiological effects of these drugs when the mutation is
studied in HEK cells or brain slices from the transgenic mice (Fig. 3.1, Fig. 4.4;
Hosie et al., 2006).
Pentobarbital’s anxiolytic effects are not particularly potent in either genotype in
our mouse strain. For percentage time spent in open arms, it seems that there
is a tendency for an overall effect of drug, but no overall effect of genotype,
which would be consistent with retained anxiolytic function of pentobarbital in
α2Q241M mutant mice. Lister (1987) demonstrated pentobarbital anxiolysis in the
elevated plus maze, but only at a dose of 30 mg/kg; it was not possible to try
this dose in our mouse strain because preliminary tests with this drug found loss
of righting reflex at 20 mg/kg. Pentobarbital was not extended to the light-dark
box paradigm due to its effects on activity, and because, the anxiolytic dose of
pentobarbital (15 mg/kg) appeared, anecdotally, to be motor-impairing in
animals of both genotypes.
Diazepam effects on wt animals are as one would expect for an anxiolytic drug
in both behavioural screens, although not all parameters in the light-dark box
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 163
are consistent with anxiolysis. As expected exploratory activity rises at low
doses (anxiolysis) and falls at higher doses (sedation). A lack of increase in
time spent in the light zone might be expected, given that THDOC also fails to
increase this parameter in wt mice. Unlike with THDOC, when examining the
‘transitions between zones’ parameter, diazepam-treated mice also fail to show
anxiolysis, only revealing sedative effects of higher doses. Methodological
differences between tests using THDOC and diazepam, such as the post-
injection time at which mice are tested (15 min and 30 min, respectively), may
account for this difference. Mice treated with diazepam-vehicle are less anxious
than the THDOC-vehicle mice (46 ± 4 vs. 36 ± 3 transitions, respectively). A
lower baseline anxiety level for the diazepam-dose response curve reduces
scope for anxiolysis to be detected for this parameter (analogous to the time
spent in the light zone).
In this study, spatiotemporal anxiety measures within the elevated plus maze
show similar anxiolytic effects for diazepam in wt and hom mice. It is unclear
why hom mice fail to display a response to diazepam in the light-dark box. The
latter test is more susceptible to any activity differences between groups being
compared, and there may be some subtle genotypic differences in activity
effects that are precluding observation of anxiolysis. One might therefore place
more faith in the elevated plus maze paradigm, in which activity effects can be
accounted for before making conclusions about anxiety. On that basis, it
appears that there is no clear effect of the mutation on diazepam sensitivity, and
the loss of THDOC sensitivity is a specific effect of the α2Q241M mutation. To
support this conclusion, it would be sensible to repeat the diazepam dose-
response analysis on the plus maze, using the wider range of doses that was
employed in the light-dark box test, to examine the anxiolytic dose-response
curve in finer detail.
One must also remember that responses to injected drugs are taking place on a
background of endogenous neurosteroids. Because the response to these
endogenous steroids is diminished in hom mice, we might also expect a
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 164
disrupted response to diazepam (see Section 3.3.3). Adding further complexity,
diazepam can increase the production of these endogenous neurosteroids, via
the ‘peripheral benzodiazepine receptor’ (now called translocator protein 18,
TSPO (Papadopoulos & Lecanu, 2009)). Interestingly, accumulating evidence
implicates this neurosteroid production as a mediator in the overall in vivo
response to benzodiazepines (e.g. midazolam’s anticonvulsant effect is partly
mediated via neurosteroid production (Dhir & Rogawski, 2012)). The lack of
diazepam-mediated anxiolysis we observe for hom mice in the light-dark box
may therefore be suggesting a role for endogenous neurosteroids in this
response. To probe whether this is the case, mice could either be pre-treated
with finasteride (to block neurosteroidogenesis), or the experiment repeated
with a benzodiazepine that does not act on TSPO, e.g. zolpidem (Trapani et al.,
1997). The former would be more challenging to implement, as measurements
would need to confirm that finasteride has successfully ablated production of
endogenous neurosteroids, but would probably be preferable to the latter (since
endogenous neurosteroids other than those produced by benzodiazepine action
at TSPO could still account for a difference in anxiolysis by zolpidem).
5.3.3. Depression may not involve neurosteroids acting at α2-type GABAA
receptors
Anxiety and depression are frequently co-morbid, both may be treated with
SSRIs and probably share common underlying patho-physiological mechanisms
(Hirschfeld, 2001; Nutt et al., 2006; Smith & Rudolph, 2012). The depressed
phenotype of α2-/- mice (Dixon et al., 2008), and involvement of neurosteroids
in depression (discussed in Section 1.4.2) may lead one to predict that hom
α2Q241M mice will be depressed as well as anxious at baseline. However, this is
not what we observe using the tail suspension test (Fig. 5.9 A). Nevertheless, a
single test cannot discount a phenotype; for example α2-/- mice were judged
‘depressed’ in most, but not all behavioural tests tried (Vollenweider et al.,
2011). Furthermore, heterozygous α2 knockout (α2+/-) mice often showed a
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 165
phenotype that was absent in homozygous knockouts (α2-/-), possibly due to
(undefined) compensatory changes in α2-/- mice that reduce depression which
are not present in the α2+/- mice. Perhaps the same is true of α2Q241M knock-in
mice; future experiments should therefore not only consider alternative tests for
depression, but also include het mice. Alternative screens for depression
include ‘novelty suppressed feeding’ (Dulawa & Hen, 2005), forced swim
(Porsolt et al., 1978) and sucrose-preference (Papp et al., 1991) tests. If
α2Q241M mice fail to show a phenotype in these tests, anxiety and depression
may well be less tightly coupled than currently thought, with anxiety more
closely linked with neurosteroid regulation of α2-type GABAA receptors.
Consistent with anxiety and depression being dissociable phenomena, GABAB
receptor knockout mice are more anxious but less depressed than wt
counterparts (Mombereau et al., 2004; Mombereau et al., 2005).
Surprisingly, THDOC does not appear to be antidepressant in wt mice in our
study (Fig. 5.9 B). We employed a tail suspension procedure that has been
previously validated (Can et al., 2012); nevertheless, unknown differences in
variables between labs can affect success of behavioural tests, so future
experiments will also include use of a known antidepressant, such as the
tricyclic agent desipramine, as a positive control. Furthermore, antidepressant
efficacy in the tail suspension test can vary widely with mouse strain (van der
Heyden et al., 1987) – perhaps the α2Q241M mutation could be tried in an
alternative background strain to C57BL/6J, in case there is some interaction of
strain with antidepressant efficacy of THDOC. Unsurprisingly on the basis of
results in other tests, THDOC increases immobility in hom mice in the tail
suspension test, probably representing the ataxic effects of the drug (rather
than an induction of behavioural despair).
Interestingly, although allopregnanolone levels are reduced in depressed
individuals, THDOC levels are increased; antidepressant SSRIs increase
allopregnanolone and decrease THDOC (Romeo et al., 1998; Uzunova et al.,
1998; Strohle et al., 1999; Strohle et al., 2000). Furthermore, whilst
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 166
antidepressant effects of progesterone and allopregnanolone have been
extensively verified in animal models (Khisti et al., 2000; e.g. Hirani et al., 2002;
Shirayama et al., 2011), we are unaware of any references demonstrating
antidepression with THDOC. Given that both compounds are GABAA receptor
potentiators, it is not clear why allopregnanolone could be antidepressant whilst
THDOC is not, but perhaps there are subtle differences in their action in vivo,
such as differential effects on non-GABAA receptor targets, that account for this
discrepancy. Furthermore, the sulphated neurosteroids DHEA-sulphate and
pregnenolone-sulphate, both negative allosteric modulators of GABAA
receptors, also demonstrate antidepressant efficacy in the tail suspension test,
probably via actions at sigma receptors (Dhir & Kulkarni, 2008). It would
therefore seem that antidepressant action by neurosteroids is not restricted to
GABAA receptor potentiation.
5.3.4. Exploring the mechanism underlying THDOC’s ataxic effects in α2Q241M
mice
Several mechanisms could account for the reduced activity of THDOC-injected
hom mice in the plus maze, light-dark box and tail suspension tests, including
sedation/sleeping, freezing or impaired motor performance. Freezing is a
normal rodent fear response; however, given that 20 mg/kg THDOC was
activity-reducing but also anxiolytic in hom mice on the elevated plus maze,
fear-induced freezing is unlikely to explain the reduced mobility. Sedation could
be studied in several ways, including loss-of-righting reflex tests or by screening
changes in electro-encephalogram (EEG) on THDOC injection; the latter may
be more relevant to this work because the marked activity reductions are
occurring at doses below the threshold for loss-of-righting reflex. During this
project, THDOC’s motor-impairing effects were assessed using an accelerating
rotarod test (Jones & Roberts, 1968). The performances of untrained wt and
hom mice were similar in this procedure (Fig. 5.10 A), suggesting the α2Q241M
mutation has no effect on motor function. After training, mice achieve a better
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 167
and more consistent performance on the rod; injection of vehicle or 10 mg/kg
THDOC did not affect performance, whilst a similar degree of impairment was
observed for both wt and hom animals after a 20 mg/kg dose (Fig. 5.10 B). We
cannot rule out a genotypic difference at other doses of THDOC, but our data
suggest that differences in motor impairment do not account for the differences
in activity effects of THDOC.
There are some important differences between the rotarod test and the other
tests described in this chapter: mice are habituated to handling during the
course of their training (cf. other tests, where the mouse is only handled for
routine husbandry and ear notching), and are effectively being forced to walk, in
order to avoid falling off the rod (cf. other tests, where mice have a free choice
over whether or not to move). It is difficult to say whether the different natures of
these tests may affect sensitivity of test subjects to THDOC’s ataxic effects, but
we can at least be sure that there are no issues of tolerance because all mice
are drug naïve. Even though different tests do appear to have slight differences
in dose-responses (e.g. significant anxiolysis for wt mice is achieved at 10
mg/kg THDOC in the plus maze, but only at 20 mg/kg in the light-dark box), the
activity effects in hom mice appear to be consistent across tests. If we therefore
consider activity effects to be similarly dose-sensitive in the rotarod, then we
have further support that motor impairment does not explain the activity effects:
10 mg/kg THDOC does not impair hom mice on the rotarod, but significantly
reduces activity in the light-dark box and elevated plus maze.
To corroborate the above results we could employ alternative tests for motor
impairment, such as the inverted screen test (Coughenour et al., 1977), which
has been argued a more sensitive test to motor impairing effects of drug.
Perhaps a wider range of doses of THDOC should be employed on both tests,
to ensure there are no genotypic differences in motor impairment at other drug
concentrations.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 168
A further explanation for activity effects is a hypersedative effect of THDOC in
hom mice; several possible mechanisms could account for increased sedation.
If, for example, the α1-type GABAA receptors are involved in sedative effects of
neurosteroids, as they are for benzodiazepines (Rudolph et al., 1999),
increased sedation could result if there is an increased activity of these subunits
in hom α2Q241M knock-ins. We can rule out a change in the α1 subunit at the
level of protein expression with our Western blot and immunofluorescence work,
but these approaches give no indication of the activity of these ion channels.
The faster decay kinetics for control sIPSCs recorded from CA1 pyramidal cells,
dentate gyrus granule cells and nucleus accumbens medium spiny neurons
(Table 4.1) may well be consistent with an increased contribution of α1-type
GABAA receptors to these sIPSCs (an increased proportion of α1 vs. α2
subunits has been shown to reduce decay time of mIPSCs and evoked IPSCs
(Okada et al., 2000)). We also cannot rule out a mechanism whereby hom mice
have a greater level of neurosteroids at baseline, such that after THDOC
injection, brain neurosteroid levels may be sufficient to pass the sedative
threshold in hom, but not wt mice. Additional experiments are needed to
address the validity of these proposals.
Sedation, as measured by a reduced locomotor activity, is often taken as an
indicator of hypnosis, but using EEG fingerprints, it is possible to dissociate
hypnosis from sedation (Mohler, 2006a). Future experiments probing the
mechanism underlying THDOC’s hyper-sedative effects will therefore utilise
EEG studies. If the mutation has led to an enhanced sensitivity to the hypnotic
effects of neurosteroids, one may predict a reduced sleep latency following
neurosteroid injection and increases in sleep vs. waking time. Potentiation of
GABAergic transmission by diazepam has also been shown to have number of
effects on rodent EEG patterns, including a rise in theta band (6-11 Hz) power
during rapid eye movement (REM) sleep, decreased delta band (0.75-4 Hz)
power in non-REM sleep and a reduced theta peak frequency during waking
(Tobler et al., 2001; Kopp et al., 2004). These investigators found that the
sleep-state EEG changes are attenuated in benzodiazepine-insensitive α2H101R
knock-in mice, but not in α1H101R mice, suggesting hypnotic effects may relate to
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 169
benzodiazepine function at α2-, rather than α1-type GABAA receptors. It would
therefore be interesting to perform similar experiments to probe the effects of
α2Q241M mutation on EEG patterns under basal conditions, and after injection of
neurosteroids.
5.4. Conclusions
1. Endogenous neurosteroids present at baseline have tonic anxiolytic
effects, but probably not antidepressant effects, via α2-type GABAA
receptors.
2. Anxiolytic effects of systemically-applied neurosteroids are mediated, at
least partly via α2-type GABAA receptors. Neurosteroids targeting the α2-
type GABAA receptor may therefore represent an appropriate treatment
strategy for anxiety disorders.
3. The α2Q241M mutation does not abolish anxiolytic effects of other GABAA
receptor potentiators, pentobarbital or diazepam. However, disrupted
diazepam anxiolysis in the light-dark box suggest that neurosteroid
responses on α2 subunits may also be required for benzodiazepine
anxiolysis.
4. Systemically applied THDOC is not antidepressant in the tail-suspension
test, and therefore may have distinct roles in anxiety and depression
compared to the related neurosteroid allopregnanolone.
Chapter 5: Screening anxiety and depression phenotypes of α2Q241M knock-ins 170
5. The ataxic effect of injected THDOC in α2Q241M mutant mice does not
derive from a hypersensitivity to motor impairing effects of this
compound.
Chapter 6: General Discussion 171
Chapter 6: General Discussion
6.1. Endogenous neurosteroids: functions at α2-type GABAA receptors
6.1.1. Anxiolysis but not anti-depression in unperturbed state
Previous work has implicated signalling through α2-type GABAA receptors as
being an important component of anxiety circuitry and anxiolysis (Low et al.,
2000; Dixon et al., 2008), and suggested that neurosteroids may act as
endogenous anxiolytics (Crawley et al., 1986; Purdy et al., 1991; Wieland et al.,
1991; Barbaccia et al., 1998). Our work provides the first direct demonstration
of a link between the two: that the endogenous anxiolytic function of
neurosteroids depends on potentiation of signalling through α2-type GABAA
receptors. Lacking this function, homozygous α2Q241M mice have a more
anxious phenotype at baseline than their wild-type littermates.
Anxiety and depression are frequently comorbid, and so these disorders may
have common underlying mechanisms (Hirschfeld, 2001; Nutt et al., 2006).
Indeed depression has also been associated with reduced GABAergic signalling
(Luscher et al., 2011b; Smith & Rudolph, 2012), possibly at α2-type GABAA
receptors (Vollenweider et al., 2011). Further, there are proposals that
antidepressants achieve their effects by restoring a deficit in allopregnanolone
levels (Romeo et al., 1998; Uzunova et al., 1998; Strohle et al., 1999; Strohle et
al., 2000). One may therefore predict a depressed phenotype to be imparted by
knocking-in the α2Q241M mutation. However, we do not observe such a
phenotype, using just one test, the tail suspension test. It is therefore possible
that anxiety and depression are less tightly coupled than currently thought, with
depression less closely linked with neurosteroid regulation of α2-type GABAA
receptors.
Chapter 6: General Discussion 172
6.1.2. α2Q241M knock-in has no effect on receptor expression
Fluctuating levels of endogenous neurosteroids can influence the expression of
GABAA receptor subunits. For example, concomitant with the allopregnanolone
level rises in the estrus cycle of mice, hippocampal expression of δ subunits
increases and that of γ2 subunits is decreased (Maguire et al., 2005). However,
we have found no change in expression of GABAA receptor α subunits in our
male α2Q241M knock-ins. This would suggest that basal neurosteroid potentiation
at α2-subunits in male mice has little influence on the expression of GABAA
receptor subunits.
6.1.3. Contribution of α2-type GABAA receptors to phasic and tonic transmission
On the basis of the effects of the α2Q241M mutation in recombinant expression
systems, one would predict that GABAergic neurotransmission will respond
normally to benzodiazepines, but have an impaired response to neurosteroids
(when currents are carried by α2-type GABAA receptors). Our recordings from
three neuronal cell types appear entirely consistent with these predictions.
Synaptic currents in wild-type animals respond to diazepam (500 nM) and
THDOC (100 nM) with a similar prolongation of decay time (i.e. enhanced
inhibitory neurotransmission). In homozygous α2Q241M mice, however, only
IPSC prolongation by diazepam is normal, and the response to THDOC is
diminished (CA1 PCs) or ablated (NAcc MSNs and DG GCs). We therefore
propose that α2-type GABAA receptors contribute significantly to the synaptic
events in these cell types, and that the residual response of CA1 PCs to
THDOC probably represents a ~50% contribution from the unaltered α subunit
isoforms that are also expressed in these cells (α1, α3, α4 and α5). This result
was to be expected given that α2 subunits play a role in synaptic targeting of
GABAA receptors (Tretter et al., 2008; Wu et al., 2012) and are strongly
Chapter 6: General Discussion 173
expressed in these cells (Fig. 3.4; Sperk et al., 1997; Pirker et al., 2000;
Sieghart & Sperk, 2002; Mohler, 2006a).
The results we observe with tonic currents were perhaps less expected.
THDOC increased the amplitude of the tonic currents recorded from DG GCs of
wild-type animals; there was no such response in cells from homozygous
α2Q241M mice, indicating a requirement for potentiation at the α2 subunit. We
believe this represents the first direct support for GABAA receptors containing
the α2 subunit playing a role in tonic currents. This is particularly interesting
given that the α2 subunit has been proposed to mostly assemble as α2β2/3γ2
receptors (McKernan & Whiting, 1996; Sieghart & Sperk, 2002; Mohler, 2006a),
and the α2 subunit plays a role in directing receptors to synaptic, rather than
extrasynaptic sites (Tretter et al., 2008; Wu et al., 2012). Such a dual role for
α2-type GABAA receptors is not unreasonable; indeed, α5βnγ2 receptor
combinations contribute to both synaptic and tonic currents in hippocampal
pyramidal cells (Banks et al., 1998; Caraiscos et al., 2004; Prenosil et al.,
2006). Interestingly, however, the synaptic events – but not tonic currents – in
our wild-type DG GC recordings show a response to diazepam. This difference
in pharmacological profile would implicate α2βnγ2 receptors in the synaptic
currents, and γ-subunit-lacking receptor combinations in the tonic currents
(Pritchett et al., 1989). We therefore propose that the THDOC-sensitive tonic
current we observe in DG GCs may instead involve receptors of combination
α2βn or α2βnδ.
6.2. The therapeutic potential of neurosteroids in anxiety and depression
6.2.1. THDOC is not antidepressant
Depression has been proposed to result from a deficiency in neurosteroids,
because allopregnanolone levels are diminished in the plasma and CSF of
depressed patients, and this deficit is ameliorated by treatment with a range of
Chapter 6: General Discussion 174
antidepressants (Romeo et al., 1998; Uzunova et al., 1998; Strohle et al., 1999;
Strohle et al., 2000). Indeed, selective-serotonin reuptake inhibitor (SSRI)
antidepressants rapidly increase allopregnanolone production in rodent models
(Uzunov et al., 1996). Consistent with neurosteroids being the mediators of
antidepression, antidepressant effects are evident with progesterone and
allopregnanolone in animal models (e.g. Khisti et al., 2000; Hirani et al., 2002;
Shirayama et al., 2011). There are, however, a number of problems with this
proposal:
1. Humans require chronic SSRI treatment before a benefit is observed,
whilst the antidepressant effects are immediate in rodents (see
discussion by Cryan et al., 2005). Interestingly, chronic fluoxetine
administration in rats actually decreases the plasma and brain
concentrations of THDOC and allopregnanolone (Serra et al., 2002).
2. Other non-SSRI antidepressants, such as imipramine, are clinically
effective, but appear not to act by influencing neurosteroid metabolism
(Uzunov et al., 1996; Griffin & Mellon, 1999).
3. The changes in THDOC levels oppose those of allopregnanolone in
depressed patients (in that THDOC levels increase in depression and are
decreased by antidepressants (Romeo et al., 1998; Uzunova et al., 1998;
Strohle et al., 1999; Strohle et al., 2000)).
4. We were unable to demonstrate an effect of THDOC in the tail
suspension test – i.e. THDOC appears not to act as an antidepressant.
Points 3 and 4 are particularly troubling when considering the suggestions that
the antidepressant function of neurosteroids occurs by enhancing GABAergic
signalling (e.g. Khisti et al., 2000). Both allopregnanolone and THDOC are
positive allosteric modulators of GABAA receptors, so it is unclear why the
increased THDOC in depressed patients would not have an antidepressant
effect. Perhaps there are subtle differences between the action of
allopregnanolone and THDOC in vivo, such as differential effects on non-
GABAA receptor targets, allowing only the former to be antidepressant.
Interestingly, sulphated steroids, which are negative allosteric modulators of
GABAA receptors, are also antidepressant in the mouse tail-suspension test,
Chapter 6: General Discussion 175
probably via actions at sigma receptors (Dhir & Kulkarni, 2008). Furthermore,
benzodiazepines – classical potentiators of GABAergic transmission – are not
generally antidepressant; only alprazolam has demonstrated any efficacy in
depression (Laakmann et al., 1996). The antidepressant action of neurosteroids
is therefore incompletely understood, and is probably not restricted to GABAA
receptor potentiation.
6.2.2. α2-type GABAA receptors are an appropriate target for neurosteroid
anxiolytics
The anxiolytic response to THDOC injection in both elevated plus maze and
light-dark box tests is disrupted by the α2Q241M knock-in. These results are
consistent with our hypothesis that α2-type GABAA receptors will be mediators
of neurosteroid anxiolysis. We propose that, analogous to benzodiazepines, the
sedative effects of neurosteroids will involve the α1 isoform. If this is the case,
subunit selective neurosteroid analogues (with efficacy at α2 and no action at
α1 subunits) would be non-sedative anxiolytics – opening an alternative avenue
for drug design to the selective benzodiazepine approach.
6.3. Remaining questions and future work
6.3.1. Screening for compensatory changes in α2Q241M mice
Our approach to screening for compensatory alterations in α2Q241M mice was to
focus on the total protein expression levels for GABAA subunits α1-α5, as likely
candidates for compensatory changes. Whilst our Western blot and
immunofluorescence data allow us to conclude that the phenotypes of this
transgenic mouse are not a result of large changes in expression of GABAA
receptors, they do not rule out changes in the activity of their ion channels.
Chapter 6: General Discussion 176
Indeed, baseline sIPSCs in knock-in mice decay faster than those of wild-type
littermates, reducing the charge transfer per synaptic event, which may indicate
some compensatory response to the mutation. However, there are numerous
observations that would argue against compensatory changes. Firstly, the
amplitude and frequency of sIPSCs is unchanged, indicating no alteration in the
number of GABAA receptors expressed at the synaptic site, and/or a lack of
alteration in presynaptic GABA release. Secondly, the sIPSC decay time
prolongation by 500 nM diazepam is undiminished by the α2Q241M mutation in
DG GCs and NAcc MSNs, further indicating an unaltered function of synaptic
GABAA receptors. Finally, tonic current amplitudes are unchanged in knock-in
mice, suggesting unchanged activity of extrasynaptic GABAA receptors. We
propose that the difference in baseline sIPSC decay times can instead be
explained by a loss of sensitivity to endogenous neurosteroids within the brain
slice. Indeed, other investigators have supported a role for endogenous
neurosteroids in modulating baseline IPSC decay times in brain slice recordings
(Puia et al., 2003).
It would be surprising if loss of neurosteroid function at α2-type GABAA
receptors does not induce some sort of compensatory change in the neuronal
network, especially since the basal inhibitory synaptic tone has been reduced in
the knock-in mice. We therefore propose that future work should use a
microarray or proteomic approach to screen for differences in expression
patterns in α2Q241M vs. wild-type mice. Furthermore, other avenues for
compensation should be considered, particularly screening for alterations in
neurosteroid levels. Since baseline inhibitory GABAergic neurotransmission in
the hippocampus is defective in the α2Q241M mutant mice, one may predict a
heightened activity of the HPA axis in these animals (see Section 1.3.1). If this
is the case, there would also be elevated production of neurosteroids and other
steroids in the adrenal cortex. Indeed the heightened plasma and CSF levels of
THDOC found in depressed patients may be a consequence of the HPA axis
hyperactivity in these individuals (Shen et al., 2010). Future work will therefore
compare the brain neurosteroid profile of α2Q241M mice with their wild-type
counterparts.
Chapter 6: General Discussion 177
6.3.2. Understanding the activity effects
Several mechanisms could account for the reduced locomotor activity observed
in knock-in mice after THDOC injection, including sedation/sleeping, freezing or
impaired motor performance. Our rotarod data suggest that α2Q241M mice are
not hyper-sensitive to the motor-impairing effects of THDOC. Future
experiments probing the mechanism underlying the hyper-sedative effects will
utilise EEG studies.
6.3.3. Future work with α2Q241M mutant mice
Would neurosteroid therapy be addictive?
The long-term use of benzodiazepines is precluded by the development of
tolerance and addiction (File, 1985; O'Brien, 2005). Would a neurosteroid-
based therapy suffer the same problems? Some investigators found that no
tolerance developed to the anticonvulsant effects of pregnenolone or the
neurosteroid analogue, ganaxolone, in rodents (Kokate et al., 1998; Reddy &
Rogawski, 2000), and ganaxolone has been used in some epileptic patients for
a number of years (Nohria & Giller, 2007). However, other investigators have
argued that tolerance does develop in response to chronic rises in neurosteroid
levels (see review by Turkmen et al., 2011). Furthermore, disorders such as
catamenial epilepsy, post-partum depression and premenstrual dysphoric
disorder have all been suggested to represent a withdrawal from high
endogenous neurosteroid levels (Smith et al., 1998; Bloch et al., 2000; Beckley
& Finn, 2007; Turkmen et al., 2011), which would suggest that withdrawal would
develop if neurosteroid therapies were abruptly terminated.
The above observations do not necessarily suggest that neurosteroid therapies
would be addictive, and a key question is whether they would have rewarding
Chapter 6: General Discussion 178
properties. The mechanisms underlying benzodiazepine reward are a subject of
debate. Encouragingly, there have been suggestions that α1-type GABAA
receptors in the ventral tegmental area are key to benzodiazepine addiction,
and that α2-selective compounds will have low addiction liability (Tan et al.,
2010; Tan et al., 2011). If neurosteroids act in a similar way, then α2-targeting
neurosteroids may prove to be non-addictive anxiolytics. However, α2 and α3-
type GABAA receptors have recently been implicated in mediating some of the
rewarding effects of benzodiazepines (which reduce the threshold for
intracranial self-stimulation (ICSS) in wild-type but not α2H101R or α3H126R
(benzodiazepine-insensitive knock-in) mice (Reynolds et al., 2012)). More work
is therefore required to further elucidate the roles of GABAergic signalling in the
mesolimbic dopamine pathway in addiction. It would be interesting to probe
whether neurosteroids have similar rewarding effects in an ICSS test, and
whether this is altered in α2Q241M mice.
Neurosteroids in relief of chronic pain?
Whilst many mechanisms underlie chronic pain (see review by Zeilhofer, 2008),
a common emerging theme involves reduced inhibitory neurotransmission in the
spinal cord dorsal horn. Potentiating inhibition via α2 subunit-containing GABAA
receptors may be an appropriate therapeutic strategy for chronic pain
syndromes (Jasmin et al., 2003; Zeilhofer et al., 2012). Furthermore, in models
of hyperalgesia, endogenously-synthesised and applied neurosteroids are
analgesic (Winter et al., 2003; Poisbeau et al., 2005). Using our α2Q241M mice,
we can probe whether signalling through α2-type GABAA receptors by
neurosteroids plays a role in pain physiology, and determine if α2-selective
neurosteroids would form an appropriate treatment strategy for neuropathic or
inflammatory pain.
Chapter 6: General Discussion 179
6.4. Concluding statement
Neurosteroids, such as allopregnanolone and THDOC, are important
endogenous modulators of GABAA receptors. They are involved in numerous
physiological processes, and are linked to several central nervous system
disorders, including depression and anxiety. Their effects in animal models
suggest they could be useful therapeutic agents, for example in anxiety, stress
and mood disorders. We have used α2Q241M knock-in mice to demonstrate that
neurosteroid-mediated anxiolysis occurs via potentiation at α2 subunit-
containing receptors. This not only informs us as to one of the key physiological
functions of endogenous neurosteroids, but also identifies the α2 isoform as an
appropriate target for generating receptor subtype-selective neurosteroid
therapeutics for anxiety disorders.
Using our knock-in strain, we have also revealed new information with regards
to the normal physiological roles of neurosteroids and GABAA receptors. We
propose that neurosteroid potentiation at α2-subunits has little influence on the
expression of GABAA receptors in male mice. Our electrophysiological
characterization demonstrated a role for α2-type GABAA receptors not only in
synaptic transmission, but also in mediating tonic currents in the dentate gyrus.
References 180
References
Akk G, Bracamontes J & Steinbach JH. (2001). Pregnenolone sulfate block of GABAA
receptors: mechanism and involvement of a residue in the M2 region of the
subunit. J Physiol 532, 673-684.
Amato A, Connolly CN, Moss SJ & Smart TG. (1999). Modulation of neuronal and
recombinant GABAA receptors by redox reagents. J Physiol 517 ( Pt 1), 35-50.
Arancibia-Carcamo IL & Kittler JT. (2009). Regulation of GABAA receptor membrane
trafficking and synaptic localization. Pharmacol Ther 123, 17-31.
Arancibia-Carcamo IL, Yuen EY, Muir J, Lumb MJ, Michels G, Saliba RS, Smart TG,
Yan Z, Kittler JT & Moss SJ. (2009). Ubiquitin-dependent lysosomal targeting of
GABAA receptors regulates neuronal inhibition. Proc Natl Acad Sci U S A 106,
17552-17557.
Atack JR. (2008). GABAA receptor subtype-selective efficacy: TPA023, an 2/ 3
selective non-sedating anxiolytic and 5IA, an 5 selective cognition enhancer.
CNS Neurosci Ther 14, 25-35.
Atack JR. (2010). GABAA receptor 2/ 3 subtype-selective modulators as potential
nonsedating anxiolytics. Curr Top Behav Neurosci 2, 331-360.
Atack JR, Hutson PH, Collinson N, Marshall G, Bentley G, Moyes C, Cook SM, Collins
I, Wafford K, McKernan RM & Dawson GR. (2005). Anxiogenic properties of an
inverse agonist selective for 3 subunit-containing GABAA receptors. Br J
Pharmacol 144, 357-366.
Atack JR, Wafford KA, Street LJ, Dawson GR, Tye S, Van Laere K, Bormans G,
Sanabria-Bohorquez SM, De Lepeleire I, de Hoon JN, Van Hecken A, Burns
HD, McKernan RM, Murphy MG & Hargreaves RJ. (2011). MRK-409 (MK-
0343), a GABAA receptor subtype-selective partial agonist, is a non-sedating
anxiolytic in preclinical species but causes sedation in humans. J
Psychopharmacol 25, 314-328.
Bai D, Zhu G, Pennefather P, Jackson MF, MacDonald JF & Orser BA. (2001). Distinct
functional and pharmacological properties of tonic and quantal inhibitory
postsynaptic currents mediated by -aminobutyric acidA receptors in
hippocampal neurons. Mol Pharmacol 59, 814-824.
References 181
Bailey DW. (1982). How pure are inbred strains of mice. Immunology Today 3, 210-
214.
Banks MI, Li TB & Pearce RA. (1998). The synaptic basis of GABAA,slow. J Neurosci 18,
1305-1317.
Bannerman DM, Rawlins JN, McHugh SB, Deacon RM, Yee BK, Bast T, Zhang WN,
Pothuizen HH & Feldon J. (2004). Regional dissociations within the
hippocampus--memory and anxiety. Neurosci Biobehav Rev 28, 273-283.
Barbaccia ML, Colombo G, Affricano D, Carai MA, Vacca G, Melis S, Purdy RH &
Gessa GL. (2002). GABAB receptor-mediated increase of neurosteroids by -
hydroxybutyric acid. Neuropharmacology 42, 782-791.
Barbaccia ML, Concas A, Serra M & Biggio G. (1998). Stress and neurosteroids in
adult and aged rats. Exp Gerontol 33, 697-712.
Barbaccia ML, Roscetti G, Trabucchi M, Purdy RH, Mostallino MC, Perra C, Concas A
& Biggio G. (1996). Isoniazid-induced inhibition of GABAergic transmission
enhances neurosteroid content in the rat brain. Neuropharmacology 35, 1299-
1305.
Barnard EA, Skolnick P, Olsen RW, Mohler H, Sieghart W, Biggio G, Braestrup C,
Bateson AN & Langer SZ. (1998). International Union of Pharmacology. XV.
Subtypes of γ-aminobutyric acidA receptors: classification on the basis of
subunit structure and receptor function. Pharmacol Rev 50, 291-313.
Beckley EH & Finn DA. (2007). Inhibition of progesterone metabolism mimics the effect
of progesterone withdrawal on forced swim test immobility. Pharmacol Biochem
Behav 87, 412-419.
Beckley EH, Scibelli AC & Finn DA. (2011). Progesterone receptor antagonist CDB-
4124 increases depression-like behavior in mice without affecting locomotor
ability. Psychoneuroendocrinology 36, 824-833.
Bedford FK, Kittler JT, Muller E, Thomas P, Uren JM, Merlo D, Wisden W, Triller A,
Smart TG & Moss SJ. (2001). GABAA receptor cell surface number and subunit
stability are regulated by the ubiquitin-like protein Plic-1. Nat Neurosci 4, 908-
916.
Belelli D, Casula A, Ling A & Lambert JJ. (2002). The influence of subunit composition
on the interaction of neurosteroids with GABAA receptors. Neuropharmacology
43, 651-661.
References 182
Belelli D & Herd MB. (2003). The contraceptive agent Provera enhances GABAA
receptor-mediated inhibitory neurotransmission in the rat hippocampus:
evidence for endogenous neurosteroids? J Neurosci 23, 10013-10020.
Belelli D, Herd MB, Mitchell EA, Peden DR, Vardy AW, Gentet L & Lambert JJ. (2006).
Neuroactive steroids and inhibitory neurotransmission: Mechanisms of action
and physiological relevance. Neuroscience 138, 821-829.
Ben-Ari Y. (2002). Excitatory actions of GABA during development: the nature of the
nurture. Nat Rev Neurosci 3, 728-739.
Benjamini Y, Drai D, Elmer G, Kafkafi N & Golani I. (2001). Controlling the false
discovery rate in behavior genetics research. Behav Brain Res 125, 279-284.
Benke D, Fritschy JM, Trzeciak A, Bannwarth W & Mohler H. (1994). Distribution,
prevalence, and drug binding profile of -aminobutyric acid type A receptor
subtypes differing in the -subunit variant. J Biol Chem 269, 27100-27107.
Benke D, Mertens S, Trzeciak A, Gillessen D & Mohler H. (1991). GABAA receptors
display association of 2-subunit with 1- and 2/3-subunits. J Biol Chem 266,
4478-4483.
Benson JA, Low K, Keist R, Mohler H & Rudolph U. (1998). Pharmacology of
recombinant -aminobutyric acidA receptors rendered diazepam-insensitive by
point-mutated subunits. FEBS Lett 431, 400-404.
Bernardi F, Salvestroni C, Casarosa E, Nappi RE, Lanzone A, Luisi S, Purdy RH,
Petraglia F & Genazzani AR. (1998). Aging is associated with changes in
allopregnanolone concentrations in brain, endocrine glands and serum in male
rats. Eur J Endocrinol 138, 316-321.
Bianchi L, Ballini C, Colivicchi MA, Della Corte L, Giovannini MG & Pepeu G. (2003).
Investigation on acetylcholine, aspartate, glutamate and GABA extracellular
levels from ventral hippocampus during repeated exploratory activity in the rat.
Neurochem Res 28, 565-573.
Bianchi MT & Macdonald RL. (2003). Neurosteroids shift partial agonist activation of
GABAA receptor channels from low- to high-efficacy gating patterns. J Neurosci
23, 10934-10943.
Biedermann K & Schoch P. (1995). Do neuroactive steroids cause fatigue in
pregnancy? Eur J Obstet Gynecol Reprod Biol 58, 15-18.
Bixo M, Andersson A, Winblad B, Purdy RH & Backstrom T. (1997). Progesterone, 5 -
pregnane-3,20-dione and 3 -hydroxy-5 -pregnane-20-one in specific regions
References 183
of the human female brain in different endocrine states. Brain Res 764, 173-
178.
Blanchard DC, Blanchard RJ, Tom P & Rodgers RJ. (1990). Diazepam changes risk
assessment in an anxiety/defense test battery. Psychopharmacology (Berl) 101,
511-518.
Bloch M, Schmidt PJ, Danaceau M, Murphy J, Nieman L & Rubinow DR. (2000).
Effects of gonadal steroids in women with a history of postpartum depression.
Am J Psychiatry 157, 924-930.
Blumstein LK & Crawley JN. (1983). Further characterization of a simple, automated
exploratory model for the anxiolytic effects of benzodiazepines. Pharmacol
Biochem Behav 18, 37-40.
Bracamontes JR & Steinbach JH. (2009). Steroid interaction with a single potentiating
site is sufficient to modulate GABA-A receptor function. Mol Pharmacol 75, 973-
981.
Brandon N, Jovanovic J & Moss S. (2002). Multiple roles of protein kinases in the
modulation of -aminobutyric acidA receptor function and cell surface
expression. Pharmacol Ther 94, 113-122.
Brickley SG, Cull-Candy SG & Farrant M. (1999). Single-channel properties of synaptic
and extrasynaptic GABAA receptors suggest differential targeting of receptor
subtypes. J Neurosci 19, 2960-2973.
Brickley SG, Revilla V, Cull-Candy SG, Wisden W & Farrant M. (2001). Adaptive
regulation of neuronal excitability by a voltage-independent potassium
conductance. Nature 409, 88-92.
Brown ES, Rush AJ & McEwen BS. (1999). Hippocampal remodeling and damage by
corticosteroids: implications for mood disorders. Neuropsychopharmacology 21,
474-484.
Brown N, Kerby J, Bonnert TP, Whiting PJ & Wafford KA. (2002). Pharmacological
characterization of a novel cell line expressing human 4 3 GABAA receptors.
Br J Pharmacol 136, 965-974.
Brussaard AB, Kits KS, Baker RE, Willems WP, Leyting-Vermeulen JW, Voorn P, Smit
AB, Bicknell RJ & Herbison AE. (1997). Plasticity in fast synaptic inhibition of
adult oxytocin neurons caused by switch in GABAA receptor subunit expression.
Neuron 19, 1103-1114.
References 184
Brussaard AB, Wossink J, Lodder JC & Kits KS. (2000). Progesterone-metabolite
prevents protein kinase C-dependent modulation of -aminobutyric acid type A
receptors in oxytocin neurons. Proc Natl Acad Sci U S A 97, 3625-3630.
Can A, Dao DT, Terrillion CE, Piantadosi SC, Bhat S & Gould TD. (2012). The tail
suspension test. J Vis Exp, e3769.
Caraiscos VB, Elliott EM, You-Ten KE, Cheng VY, Belelli D, Newell JG, Jackson MF,
Lambert JJ, Rosahl TW, Wafford KA, MacDonald JF & Orser BA. (2004). Tonic
inhibition in mouse hippocampal CA1 pyramidal neurons is mediated by 5
subunit-containing γ-aminobutyric acid type A receptors. Proc Natl Acad Sci U
S A 101, 3662-3667.
Carter RB, Wood PL, Wieland S, Hawkinson JE, Belelli D, Lambert JJ, White HS, Wolf
HH, Mirsadeghi S, Tahir SH, Bolger MB, Lan NC & Gee KW. (1997).
Characterization of the anticonvulsant properties of ganaxolone (CCD 1042; 3 -
hydroxy-3 -methyl-5 -pregnan-20-one), a selective, high-affinity, steroid
modulator of the -aminobutyric acidA receptor. J Pharmacol Exp Ther 280,
1284-1295.
Chandra D, Korpi ER, Miralles CP, De Blas AL & Homanics GE. (2005). GABAA
receptor 2 subunit knockdown mice have enhanced anxiety-like behavior but
unaltered hypnotic response to benzodiazepines. BMC Neurosci 6, 30.
Chang Y, Wang R, Barot S & Weiss DS. (1996). Stoichiometry of a recombinant
GABAA receptor. J Neurosci 16, 5415-5424.
Chiang PH, Wu PY, Kuo TW, Liu YC, Chan CF, Chien TC, Cheng JK, Huang YY, Chiu
CD & Lien CC. (2012). GABA is depolarizing in hippocampal dentate granule
cells of the adolescent and adult rats. J Neurosci 32, 62-67.
Chisari M, Eisenman LN, Covey DF, Mennerick S & Zorumski CF. (2010). The sticky
issue of neurosteroids and GABAA receptors. Trends Neurosci 33, 299-306.
Concas A, Mostallino MC, Porcu P, Follesa P, Barbaccia ML, Trabucchi M, Purdy RH,
Grisenti P & Biggio G. (1998). Role of brain allopregnanolone in the plasticity of
-aminobutyric acid type A receptor in rat brain during pregnancy and after
delivery. Proc Natl Acad Sci U S A 95, 13284-13289.
Connolly CN, Kittler JT, Thomas P, Uren JM, Brandon NJ, Smart TG & Moss SJ.
(1999). Cell surface stability of -aminobutyric acid type A receptors.
Dependence on protein kinase C activity and subunit composition. J Biol Chem
274, 36565-36572.
References 185
Cooper EJ, Johnston GA & Edwards FA. (1999). Effects of a naturally occurring
neurosteroid on GABAA IPSCs during development in rat hippocampal or
cerebellar slices. J Physiol 521 Pt 2, 437-449.
Corpechot C, Collins BE, Carey MP, Tsouros A, Robel P & Fry JP. (1997). Brain
neurosteroids during the mouse oestrous cycle. Brain Res 766, 276-280.
Corringer PJ, Poitevin F, Prevost MS, Sauguet L, Delarue M & Changeux JP. (2012).
Structure and pharmacology of pentameric receptor channels: from bacteria to
brain. Structure 20, 941-956.
Coughenour LL, McLean JR & Parker RB. (1977). A new device for the rapid
measurement of impaired motor function in mice. Pharmacol Biochem Behav 6,
351-353.
Crawley JN, Glowa JR, Majewska MD & Paul SM. (1986). Anxiolytic activity of an
endogenous adrenal steroid. Brain Res 398, 382-385.
Crawley JN & Goodwin FK. (1980). Preliminary report of a simple animal behavior
model for the anxiolytic effects of benzodiazepines. Pharmacol Biochem Behav
13, 167-170.
Crestani F, Lorez M, Baer K, Essrich C, Benke D, Laurent JP, Belzung C, Fritschy JM,
Luscher B & Mohler H. (1999). Decreased GABAA-receptor clustering results in
enhanced anxiety and a bias for threat cues. Nat Neurosci 2, 833-839.
Crestani F, Low K, Keist R, Mandelli M, Mohler H & Rudolph U. (2001). Molecular
targets for the myorelaxant action of diazepam. Mol Pharmacol 59, 442-445.
Cryan JF, Mombereau C & Vassout A. (2005). The tail suspension test as a model for
assessing antidepressant activity: Review of pharmacological and genetic
studies in mice. Neurosci Biobehav Rev 29, 571-625.
de Lanerolle NC, Kim JH, Williamson A, Spencer SS, Zaveri HP, Eid T & Spencer DD.
(2003). A retrospective analysis of hippocampal pathology in human temporal
lobe epilepsy: evidence for distinctive patient subcategories. Epilepsia 44, 677-
687.
Dean B, Scarr E & McLeod M. (2005). Changes in hippocampal GABAA receptor
subunit composition in bipolar 1 disorder. Brain Res Mol Brain Res 138, 145-
155.
Dhir A & Kulkarni S. (2008). Involvement of sigma ( 1) receptors in modulating the
anti-depressant effect of neurosteroids (dehydroepiandrosterone or
pregnenolone) in mouse tail-suspension test. J Psychopharmacol 22, 691-696.
References 186
Dhir A & Rogawski MA. (2012). Role of neurosteroids in the anticonvulsant activity of
midazolam. Br J Pharmacol 165, 2684-2691.
Dias R, Sheppard WF, Fradley RL, Garrett EM, Stanley JL, Tye SJ, Goodacre S,
Lincoln RJ, Cook SM, Conley R, Hallett D, Humphries AC, Thompson SA,
Wafford KA, Street LJ, Castro JL, Whiting PJ, Rosahl TW, Atack JR, McKernan
RM, Dawson GR & Reynolds DS. (2005). Evidence for a significant role of 3-
containing GABAA receptors in mediating the anxiolytic effects of
benzodiazepines. J Neurosci 25, 10682-10688.
Dixon CI, Morris HV, Breen G, Desrivieres S, Jugurnauth S, Steiner RC, Vallada H,
Guindalini C, Laranjeira R, Messas G, Rosahl TW, Atack JR, Peden DR, Belelli
D, Lambert JJ, King SL, Schumann G & Stephens DN. (2010). Cocaine effects
on mouse incentive-learning and human addiction are linked to 2 subunit-
containing GABAA receptors. Proc Natl Acad Sci U S A 107, 2289-2294.
Dixon CI, Rosahl TW & Stephens DN. (2008). Targeted deletion of the GABRA2 gene
encoding 2-subunits of GABAA receptors facilitates performance of a
conditioned emotional response, and abolishes anxiolytic effects of
benzodiazepines and barbiturates. Pharmacol Biochem Behav 90, 1-8.
Dulawa SC & Hen R. (2005). Recent advances in animal models of chronic
antidepressant effects: The novelty-induced hypophagia test. Neurosci
Biobehav Rev 29, 771-783.
Dunham NW & Miya TS. (1957). A note on a simple apparatus for detecting
neurological deficit in rats and mice. J Am Pharm Assoc Am Pharm Assoc
(Baltim) 46, 208-209.
Earnheart JC, Schweizer C, Crestani F, Iwasato T, Itohara S, Mohler H & Luscher B.
(2007). GABAergic control of adult hippocampal neurogenesis in relation to
behavior indicative of trait anxiety and depression states. J Neurosci 27, 3845-
3854.
Essrich C, Lorez M, Benson JA, Fritschy JM & Luscher B. (1998). Postsynaptic
clustering of major GABAA receptor subtypes requires the 2 subunit and
gephyrin. Nat Neurosci 1, 563-571.
Fancsik A, Linn DM & Tasker JG. (2000). Neurosteroid modulation of GABA IPSCs is
phosphorylation dependent. J Neurosci 20, 3067-3075.
Fanselow MS & Dong HW. (2010). Are the dorsal and ventral hippocampus functionally
distinct structures? Neuron 65, 7-19.
References 187
Farisello P, Boido D, Nieus T, Medrihan L, Cesca F, Valtorta F, Baldelli P & Benfenati
F. (2012). Synaptic and extrasynaptic origin of the excitation/inhibition
imbalance in the hippocampus of synapsin I/II/III knockout mice. Cereb Cortex.
Farrant M & Nusser Z. (2005). Variations on an inhibitory theme: phasic and tonic
activation of GABAA receptors. Nat Rev Neurosci 6, 215-229.
Farrar SJ, Whiting PJ, Bonnert TP & McKernan RM. (1999). Stoichiometry of a ligand-
gated ion channel determined by fluorescence energy transfer. J Biol Chem
274, 10100-10104.
File SE. (1985). Tolerance to the behavioral actions of benzodiazepines. Neurosci
Biobehav Rev 9, 113-121.
Fischer F, Kneussel M, Tintrup H, Haverkamp S, Rauen T, Betz H & Wassle H. (2000).
Reduced synaptic clustering of GABA and glycine receptors in the retina of the
gephyrin null mutant mouse. J Comp Neurol 427, 634-648.
Fritschy JM, Benke D, Mertens S, Oertel WH, Bachi T & Mohler H. (1992). Five
subtypes of type A -aminobutyric acid receptors identified in neurons by double
and triple immunofluorescence staining with subunit-specific antibodies. Proc
Natl Acad Sci U S A 89, 6726-6730.
Fritschy JM & Brunig I. (2003). Formation and plasticity of GABAergic synapses:
physiological mechanisms and pathophysiological implications. Pharmacol Ther
98, 299-323.
Glencorse TA, Bateson AN & Darlison MG. (1992). Differential Localization of Two
Alternatively Spliced GABAA Receptor 2-Subunit mRNAs in the Chick Brain.
Eur J Neurosci 4, 271-277.
Glykys J, Mann EO & Mody I. (2008). Which GABAA receptor subunits are necessary
for tonic inhibition in the hippocampus? J Neurosci 28, 1421-1426.
Glykys J & Mody I. (2007). The main source of ambient GABA responsible for tonic
inhibition in the mouse hippocampus. J Physiol 582, 1163-1178.
Griffin LD, Gong W, Verot L & Mellon SH. (2004). Niemann-Pick type C disease
involves disrupted neurosteroidogenesis and responds to allopregnanolone. Nat
Med 10, 704-711.
Griffin LD & Mellon SH. (1999). Selective serotonin reuptake inhibitors directly alter
activity of neurosteroidogenic enzymes. Proc Natl Acad Sci U S A 96, 13512-
13517.
Gunther U, Benson J, Benke D, Fritschy JM, Reyes G, Knoflach F, Crestani F, Aguzzi
A, Arigoni M, Lang Y & et al. (1995). Benzodiazepine-insensitive mice
References 188
generated by targeted disruption of the 2 subunit gene of -aminobutyric acid
type A receptors. Proc Natl Acad Sci U S A 92, 7749-7753.
Harney SC, Frenguelli BG & Lambert JJ. (2003). Phosphorylation influences
neurosteroid modulation of synaptic GABAA receptors in rat CA1 and dentate
gyrus neurones. Neuropharmacology 45, 873-883.
Harris RA, Osterndorff-Kahanek E, Ponomarev I, Homanics GE & Blednov YA. (2011).
Testing the silence of mutations: Transcriptomic and behavioral studies of
GABAA receptor 1 and 2 subunit knock-in mice. Neurosci Lett 488, 31-35.
Harrison NL & Simmonds MA. (1984). Modulation of the GABA receptor complex by a
steroid anaesthetic. Brain Res 323, 287-292.
Harrison NL, Vicini S & Barker JL. (1987). A steroid anesthetic prolongs inhibitory
postsynaptic currents in cultured rat hippocampal neurons. J Neurosci 7, 604-
609.
Hascoet M & Bourin M. (1998). A new approach to the light/dark test procedure in
mice. Pharmacol Biochem Behav 60, 645-653.
Hasler G, van der Veen JW, Tumonis T, Meyers N, Shen J & Drevets WC. (2007).
Reduced prefrontal glutamate/glutamine and -aminobutyric acid levels in major
depression determined using proton magnetic resonance spectroscopy. Arch
Gen Psychiatry 64, 193-200.
Heimer L, Alheid GF, de Olmos JS, Groenewegen HJ, Haber SN, Harlan RE & Zahm
DS. (1997). The accumbens: beyond the core-shell dichotomy. J
Neuropsychiatry Clin Neurosci 9, 354-381.
Helms CM, Rossi DJ & Grant KA. (2012). Neurosteroid influences on sensitivity to
ethanol. Front Endocrinol (Lausanne) 3, 10.
Henkel V, Baghai TC, Eser D, Zill P, Mergl R, Zwanzger P, Schule C, Bottlender R,
Jager M, Rupprecht R, Hegerl U, Moller HJ & Bondy B. (2004). The gamma
amino butyric acid (GABA) receptor alpha-3 subunit gene polymorphism in
unipolar depressive disorder: a genetic association study. Am J Med Genet B
Neuropsychiatr Genet 126B, 82-87.
Hibbs RE & Gouaux E. (2011). Principles of activation and permeation in an anion-
selective Cys-loop receptor. Nature 474, 54-60.
Hirani K, Khisti RT & Chopde CT. (2002). Behavioral action of ethanol in Porsolt's
forced swim test: modulation by 3alpha-hydroxy-5alpha-pregnan-20-one.
Neuropharmacology 43, 1339-1350.
References 189
Hirschfeld RM. (2001). The comorbidity of major depression and anxiety disorders:
Recognition and management in primary care. Prim Care Companion J Clin
Psychiatry 3, 244-254.
Hogg S. (1996). A review of the validity and variability of the elevated plus-maze as an
animal model of anxiety. Pharmacol Biochem Behav 54, 21-30.
Honig A, Bartlett JR, Bouras N & Bridges PK. (1988). Amino acid levels in depression:
a preliminary investigation. J Psychiatr Res 22, 159-164.
Horiuchi Y, Nakayama J, Ishiguro H, Ohtsuki T, Detera-Wadleigh SD, Toyota T,
Yamada K, Nankai M, Shibuya H, Yoshikawa T & Arinami T. (2004). Possible
association between a haplotype of the GABA-A receptor alpha 1 subunit gene
(GABRA1) and mood disorders. Biol Psychiatry 55, 40-45.
Hosie AM, Clarke L, da Silva H & Smart TG. (2009). Conserved site for neurosteroid
modulation of GABAA receptors. Neuropharmacology 56, 149-154.
Hosie AM, Dunne EL, Harvey RJ & Smart TG. (2003). Zinc-mediated inhibition of
GABAA receptors: discrete binding sites underlie subtype specificity. Nat
Neurosci 6, 362-369.
Hosie AM, Wilkins ME, da Silva HM & Smart TG. (2006). Endogenous neurosteroids
regulate GABAA receptors through two discrete transmembrane sites. Nature
444, 486-489.
Hosie AM, Wilkins ME & Smart TG. (2007). Neurosteroid binding sites on GABAA
receptors. Pharmacol Ther 116, 7-19.
Houston CM, Bright DP, Sivilotti LG, Beato M & Smart TG. (2009a). Intracellular
chloride ions regulate the time course of GABA-mediated inhibitory synaptic
transmission. J Neurosci 29, 10416-10423.
Houston CM, He Q & Smart TG. (2009b). CaMKII phosphorylation of the GABAA
receptor: receptor subtype- and synapse-specific modulation. J Physiol 587,
2115-2125.
Hu W, Zhang MY, Czeh B, Flugge G & Zhang WQ. (2010). Stress impairs GABAergic
network function in the hippocampus by activating nongenomic glucocorticoid
receptors and affecting the integrity of the parvalbumin-expressing neuronal
network. Neuropsychopharmacology 35, 1693-1707.
Huang RQ & Dillon GH. (1999). Effect of extracellular pH on GABA-activated current in
rat recombinant receptors and thin hypothalamic slices. J Neurophysiol 82,
1233-1243.
References 190
Hyman SE & Malenka RC. (2001). Addiction and the brain: the neurobiology of
compulsion and its persistence. Nat Rev Neurosci 2, 695-703.
Ichikawa S, Sawada T, Nakamura Y & Morioka H. (1974). Ovarian secretion of
pregnane compounds during the estrous cycle and pregnancy in rats.
Endocrinology 94, 1615-1620.
Item C & Sieghart W. (1994). Binding of -aminobutyric acidA receptors to tubulin. J
Neurochem 63, 1119-1125.
Jasmin L, Rabkin SD, Granato A, Boudah A & Ohara PT. (2003). Analgesia and
hyperalgesia from GABA-mediated modulation of the cerebral cortex. Nature
424, 316-320.
Jones BJ & Roberts DJ. (1968). A rotarod suitable for quantitative measurements of
motor incoordination in naive mice. Naunyn Schmiedebergs Arch Exp Pathol
Pharmakol 259, 211.
Jovanovic JN, Thomas P, Kittler JT, Smart TG & Moss SJ. (2004). Brain-derived
neurotrophic factor modulates fast synaptic inhibition by regulating GABAA
receptor phosphorylation, activity, and cell-surface stability. J Neurosci 24, 522-
530.
Kanematsu T, Yasunaga A, Mizoguchi Y, Kuratani A, Kittler JT, Jovanovic JN,
Takenaka K, Nakayama KI, Fukami K, Takenawa T, Moss SJ, Nabekura J &
Hirata M. (2006). Modulation of GABAA receptor phosphorylation and
membrane trafficking by phospholipase C-related inactive protein/protein
phosphatase 1 and 2A signaling complex underlying brain-derived neurotrophic
factor-dependent regulation of GABAergic inhibition. J Biol Chem 281, 22180-
22189.
Khisti RT, Chopde CT & Jain SP. (2000). Antidepressant-like effect of the neurosteroid
3 -hydroxy-5 -pregnan-20-one in mice forced swim test. Pharmacol Biochem
Behav 67, 137-143.
Khisti RT, Penland SN, VanDoren MJ, Grobin AC & Morrow AL. (2002). GABAergic
neurosteroid modulation of ethanol actions. World J Biol Psychiatry 3, 87-95.
Kittler JT, Chen G, Honing S, Bogdanov Y, McAinsh K, Arancibia-Carcamo IL,
Jovanovic JN, Pangalos MN, Haucke V, Yan Z & Moss SJ. (2005). Phospho-
dependent binding of the clathrin AP2 adaptor complex to GABAA receptors
regulates the efficacy of inhibitory synaptic transmission. Proc Natl Acad Sci U
S A 102, 14871-14876.
References 191
Kittler JT, Delmas P, Jovanovic JN, Brown DA, Smart TG & Moss SJ. (2000).
Constitutive endocytosis of GABAA receptors by an association with the adaptin
AP2 complex modulates inhibitory synaptic currents in hippocampal neurons. J
Neurosci 20, 7972-7977.
Kittler JT & Moss SJ. (2001). Neurotransmitter receptor trafficking and the regulation of
synaptic strength. Traffic 2, 437-448.
Kittler JT & Moss SJ. (2003). Modulation of GABAA receptor activity by phosphorylation
and receptor trafficking: implications for the efficacy of synaptic inhibition. Curr
Opin Neurobiol 13, 341-347.
Kittler JT, Rostaing P, Schiavo G, Fritschy JM, Olsen R, Triller A & Moss SJ. (2001).
The subcellular distribution of GABARAP and its ability to interact with NSF
suggest a role for this protein in the intracellular transport of GABAA receptors.
Mol Cell Neurosci 18, 13-25.
Kneussel M, Haverkamp S, Fuhrmann JC, Wang H, Wassle H, Olsen RW & Betz H.
(2000). The -aminobutyric acid type A receptor (GABAAR)-associated protein
GABARAP interacts with gephyrin but is not involved in receptor anchoring at
the synapse. Proc Natl Acad Sci U S A 97, 8594-8599.
Knuesel I, Mastrocola M, Zuellig RA, Bornhauser B, Schaub MC & Fritschy JM. (1999).
Short communication: altered synaptic clustering of GABAA receptors in mice
lacking dystrophin (mdx mice). Eur J Neurosci 11, 4457-4462.
Kofuji P, Wang JB, Moss SJ, Huganir RL & Burt DR. (1991). Generation of two forms of
the -aminobutyric acidA receptor 2-subunit in mice by alternative splicing. J
Neurochem 56, 713-715.
Kokate TG, Yamaguchi S, Pannell LK, Rajamani U, Carroll DM, Grossman AB &
Rogawski MA. (1998). Lack of anticonvulsant tolerance to the neuroactive
steroid pregnanolone in mice. J Pharmacol Exp Ther 287, 553-558.
Koksma JJ, van Kesteren RE, Rosahl TW, Zwart R, Smit AB, Luddens H & Brussaard
AB. (2003). Oxytocin regulates neurosteroid modulation of GABAA receptors in
supraoptic nucleus around parturition. J Neurosci 23, 788-797.
Kopp C, Rudolph U, Low K & Tobler I. (2004). Modulation of rhythmic brain activity by
diazepam: GABAA receptor subtype and state specificity. Proc Natl Acad Sci U
S A 101, 3674-3679.
Korpi ER, Grunder G & Luddens H. (2002). Drug interactions at GABAA receptors. Prog
Neurobiol 67, 113-159.
References 192
Kralic JE, O'Buckley TK, Khisti RT, Hodge CW, Homanics GE & Morrow AL. (2002).
GABAA receptor alpha-1 subunit deletion alters receptor subtype assembly,
pharmacological and behavioral responses to benzodiazepines and zolpidem.
Neuropharmacology 43, 685-694.
Krishnan V & Nestler EJ. (2008). The molecular neurobiology of depression. Nature
455, 894-902.
Kugaya A, Sanacora G, Verhoeff NP, Fujita M, Mason GF, Seneca NM, Bozkurt A,
Khan SA, Anand A, Degen K, Charney DS, Zoghbi SS, Baldwin RM, Seibyl JP
& Innis RB. (2003). Cerebral benzodiazepine receptors in depressed patients
measured with [123I]iomazenil SPECT. Biol Psychiatry 54, 792-799.
Laakmann G, Faltermaier-Temizel M, Bossert-Zaudig S & Baghai T. (1996). Are
benzodiazepines antidepressants? Psychopharmacology (Berl) 124, 291-292.
Lacasse JR & Leo J. (2005). Serotonin and depression: a disconnect between the
advertisements and the scientific literature. PLoS Med 2, e392.
Lambert JJ, Belelli D, Peden DR, Vardy AW & Peters JA. (2003). Neurosteroid
modulation of GABAA receptors. Prog Neurobiol 71, 67-80.
Laurie DJ, Seeburg PH & Wisden W. (1992a). The distribution of 13 GABAA receptor
subunit mRNAs in the rat brain. II. Olfactory bulb and cerebellum. J Neurosci
12, 1063-1076.
Laurie DJ, Wisden W & Seeburg PH. (1992b). The distribution of thirteen GABAA
receptor subunit mRNAs in the rat brain. III. Embryonic and postnatal
development. J Neurosci 12, 4151-4172.
Li GD, Chiara DC, Cohen JB & Olsen RW. (2009). Neurosteroids allosterically
modulate binding of the anesthetic etomidate to -aminobutyric acid type A
receptors. J Biol Chem 284, 11771-11775.
Li P, Shu HJ, Wang C, Mennerick S, Zorumski CF, Covey DF, Steinbach JH & Akk G.
(2007). Neurosteroid migration to intracellular compartments reduces steroid
concentration in the membrane and diminishes GABAA receptor potentiation. J
Physiol 584, 789-800.
Lippa A, Czobor P, Stark J, Beer B, Kostakis E, Gravielle M, Bandyopadhyay S,
Russek SJ, Gibbs TT, Farb DH & Skolnick P. (2005). Selective anxiolysis
produced by ocinaplon, a GABAA receptor modulator. Proc Natl Acad Sci U S A
102, 7380-7385.
Lister RG. (1987). The use of a plus-maze to measure anxiety in the mouse.
Psychopharmacology (Berl) 92, 180-185.
References 193
Low K, Crestani F, Keist R, Benke D, Brunig I, Benson JA, Fritschy JM, Rulicke T,
Bluethmann H, Mohler H & Rudolph U. (2000). Molecular and neuronal
substrate for the selective attenuation of anxiety. Science 290, 131-134.
Luisi S, Petraglia F, Benedetto C, Nappi RE, Bernardi F, Fadalti M, Reis FM, Luisi M &
Genazzani AR. (2000). Serum allopregnanolone levels in pregnant women:
changes during pregnancy, at delivery, and in hypertensive patients. J Clin
Endocrinol Metab 85, 2429-2433.
Luscher B, Fuchs T & Kilpatrick CL. (2011a). GABAA receptor trafficking-mediated
plasticity of inhibitory synapses. Neuron 70, 385-409.
Luscher B & Keller CA. (2004). Regulation of GABAA receptor trafficking, channel
activity, and functional plasticity of inhibitory synapses. Pharmacol Ther 102,
195-221.
Luscher B, Shen Q & Sahir N. (2011b). The GABAergic deficit hypothesis of major
depressive disorder. Mol Psychiatry 16, 383-406.
Luscher C & Ungless MA. (2006). The mechanistic classification of addictive drugs.
PLoS Med 3, e437.
MacQueen G & Frodl T. (2011). The hippocampus in major depression: evidence for
the convergence of the bench and bedside in psychiatric research? Mol
Psychiatry 16, 252-264.
Maguire J & Mody I. (2008). GABAAR plasticity during pregnancy: relevance to
postpartum depression. Neuron 59, 207-213.
Maguire JL, Stell BM, Rafizadeh M & Mody I. (2005). Ovarian cycle-linked changes in
GABAA receptors mediating tonic inhibition alter seizure susceptibility and
anxiety. Nat Neurosci 8, 797-804.
McCabe C, Cowen PJ & Harmer CJ. (2009). Neural representation of reward in
recovered depressed patients. Psychopharmacology (Berl) 205, 667-677.
McKernan RM, Rosahl TW, Reynolds DS, Sur C, Wafford KA, Atack JR, Farrar S,
Myers J, Cook G, Ferris P, Garrett L, Bristow L, Marshall G, Macaulay A, Brown
N, Howell O, Moore KW, Carling RW, Street LJ, Castro JL, Ragan CI, Dawson
GR & Whiting PJ. (2000). Sedative but not anxiolytic properties of
benzodiazepines are mediated by the GABAA receptor 1 subtype. Nat
Neurosci 3, 587-592.
McKernan RM & Whiting PJ. (1996). Which GABAA receptor subtypes really occur in
the brain? Trends Neurosci 19, 139-143.
References 194
Mellon SH, Gong W & Schonemann MD. (2008). Endogenous and synthetic
neurosteroids in treatment of Niemann-Pick Type C disease. Brain Res Rev 57,
410-420.
Mellon SH & Griffin LD. (2002). Neurosteroids: biochemistry and clinical significance.
Trends Endocrinol Metab 13, 35-43.
Mendelson WB, Martin JV, Perlis M, Wagner R, Majewska MD & Paul SM. (1987).
Sleep induction by an adrenal steroid in the rat. Psychopharmacology (Berl) 93,
226-229.
Mihalek RM, Banerjee PK, Korpi ER, Quinlan JJ, Firestone LL, Mi ZP, Lagenaur C,
Tretter V, Sieghart W, Anagnostaras SG, Sage JR, Fanselow MS, Guidotti A,
Spigelman I, Li Z, DeLorey TM, Olsen RW & Homanics GE. (1999). Attenuated
sensitivity to neuroactive steroids in -aminobutyrate type A receptor delta
subunit knockout mice. Proc Natl Acad Sci U S A 96, 12905-12910.
Miller PS & Smart TG. (2010). Binding, activation and modulation of Cys-loop
receptors. Trends Pharmacol Sci 31, 161-174.
Mizokami A, Kanematsu T, Ishibashi H, Yamaguchi T, Tanida I, Takenaka K,
Nakayama KI, Fukami K, Takenawa T, Kominami E, Moss SJ, Yamamoto T,
Nabekura J & Hirata M. (2007). Phospholipase C-related inactive protein is
involved in trafficking of 2 subunit-containing GABAA receptors to the cell
surface. J Neurosci 27, 1692-1701.
Mody I. (2001). Distinguishing between GABAA receptors responsible for tonic and
phasic conductances. Neurochem Res 26, 907-913.
Mody I & Maguire J. (2011). The reciprocal regulation of stress hormones and GABAA
receptors. Front Cell Neurosci 6, 4.
Mohler H. (2006a). GABAA receptor diversity and pharmacology. Cell Tissue Res 326,
505-516.
Mohler H. (2006b). GABAA receptors in central nervous system disease: anxiety,
epilepsy, and insomnia. J Recept Signal Transduct Res 26, 731-740.
Mohler H. (2011). The rise of a new GABA pharmacology. Neuropharmacology 60,
1042-1049.
Mombereau C, Kaupmann K, Froestl W, Sansig G, van der Putten H & Cryan JF.
(2004). Genetic and pharmacological evidence of a role for GABAB receptors in
the modulation of anxiety- and antidepressant-like behavior.
Neuropsychopharmacology 29, 1050-1062.
References 195
Mombereau C, Kaupmann K, Gassmann M, Bettler B, van der Putten H & Cryan JF.
(2005). Altered anxiety and depression-related behaviour in mice lacking
GABAB(2) receptor subunits. Neuroreport 16, 307-310.
Morris HV, Dawson GR, Reynolds DS, Atack JR, Rosahl TW & Stephens DN. (2008).
Alpha2-containing GABAA receptors are involved in mediating stimulant effects
of cocaine. Pharmacol Biochem Behav 90, 9-18.
Mortensen M & Smart TG. (2006). Extrasynaptic subunit GABAA receptors on rat
hippocampal pyramidal neurons. J Physiol 577, 841-856.
Moss SJ & Smart TG. (2001). Constructing inhibitory synapses. Nat Rev Neurosci 2,
240-250.
Mtchedlishvili Z, Sun CS, Harrison MB & Kapur J. (2003). Increased neurosteroid
sensitivity of hippocampal GABAA receptors during postnatal development.
Neuroscience 118, 655-666.
Munzel U & Hothorn LA. (2001). A unified approach to simultaneous rank test
procedures in the unbalanced one-way layout. Biometrical Journal 43, 553-569.
Nguyen PN, Billiards SS, Walker DW & Hirst JJ. (2003). Changes in 5 -pregnane
steroids and neurosteroidogenic enzyme expression in the perinatal sheep.
Pediatr Res 53, 956-964.
Nohria V & Giller E. (2007). Ganaxolone. Neurotherapeutics 4, 102-105.
Nusser Z & Mody I. (2002). Selective modulation of tonic and phasic inhibitions in
dentate gyrus granule cells. J Neurophysiol 87, 2624-2628.
Nusser Z, Sieghart W, Benke D, Fritschy JM & Somogyi P. (1996). Differential synaptic
localization of two major -aminobutyric acid type A receptor subunits on
hippocampal pyramidal cells. Proc Natl Acad Sci U S A 93, 11939-11944.
Nusser Z, Sieghart W & Somogyi P. (1998). Segregation of different GABAA receptors
to synaptic and extrasynaptic membranes of cerebellar granule cells. J
Neurosci 18, 1693-1703.
Nutt D, Argyropoulos S, Hood S & Potokar J. (2006). Generalized anxiety disorder: A
comorbid disease. Eur Neuropsychopharmacol 16 Suppl 2, S109-118.
Nyiri G, Freund TF & Somogyi P. (2001). Input-dependent synaptic targeting of 2-
subunit-containing GABAA receptors in synapses of hippocampal pyramidal
cells of the rat. Eur J Neurosci 13, 428-442.
O'Brien CP. (2005). Benzodiazepine use, abuse, and dependence. J Clin Psychiatry 66
Suppl 2, 28-33.
References 196
Okada M, Onodera K, Van Renterghem C, Sieghart W & Takahashi T. (2000).
Functional correlation of GABAA receptor subunits expression with the
properties of IPSCs in the developing thalamus. J Neurosci 20, 2202-2208.
Olsen RW & Li GD. (2011). GABAA receptors as molecular targets of general
anesthetics: identification of binding sites provides clues to allosteric
modulation. Can J Anaesth 58, 206-215.
Otis TS & Mody I. (1992). Modulation of decay kinetics and frequency of GABAA
receptor-mediated spontaneous inhibitory postsynaptic currents in hippocampal
neurons. Neuroscience 49, 13-32.
Overstreet LS, Jones MV & Westbrook GL. (2000). Slow desensitization regulates the
availability of synaptic GABAA receptors. J Neurosci 20, 7914-7921.
Pan ZH, Zhang X & Lipton SA. (2000). Redox modulation of recombinant human
GABAA receptors. Neuroscience 98, 333-338.
Papadopoulos V & Lecanu L. (2009). Translocator protein (18 kDa) TSPO: an
emerging therapeutic target in neurotrauma. Exp Neurol 219, 53-57.
Papp M, Willner P & Muscat R. (1991). An animal model of anhedonia: attenuation of
sucrose consumption and place preference conditioning by chronic
unpredictable mild stress. Psychopharmacology (Berl) 104, 255-259.
Paul SM & Purdy RH. (1992). Neuroactive steroids. FASEB J 6, 2311-2322.
Pellow S, Chopin P, File SE & Briley M. (1985). Validation of open:closed arm entries
in an elevated plus-maze as a measure of anxiety in the rat. J Neurosci
Methods 14, 149-167.
Pellow S & File SE. (1986). Anxiolytic and anxiogenic drug effects on exploratory
activity in an elevated plus-maze: a novel test of anxiety in the rat. Pharmacol
Biochem Behav 24, 525-529.
Petty F. (1994). Plasma concentrations of -aminobutyric acid (GABA) and mood
disorders: a blood test for manic depressive disease? Clin Chem 40, 296-302.
Pham X, Sun C, Chen X, van den Oord EJ, Neale MC, Kendler KS & Hettema JM.
(2009). Association study between GABA receptor genes and anxiety spectrum
disorders. Depress Anxiety 26, 998-1003.
Pirker S, Schwarzer C, Wieselthaler A, Sieghart W & Sperk G. (2000). GABAA
receptors: immunocytochemical distribution of 13 subunits in the adult rat brain.
Neuroscience 101, 815-850.
Poisbeau P, Patte-Mensah C, Keller AF, Barrot M, Breton JD, Luis-Delgado OE,
Freund-Mercier MJ, Mensah-Nyagan AG & Schlichter R. (2005). Inflammatory
References 197
pain upregulates spinal inhibition via endogenous neurosteroid production. J
Neurosci 25, 11768-11776.
Porcu P, Sogliano C, Cinus M, Purdy RH, Biggio G & Concas A. (2003). Nicotine-
induced changes in cerebrocortical neuroactive steroids and plasma
corticosterone concentrations in the rat. Pharmacol Biochem Behav 74, 683-
690.
Porsolt RD, Bertin A & Jalfre M. (1978). "Behavioural despair" in rats and mice: strain
differences and the effects of imipramine. Eur J Pharmacol 51, 291-294.
Prenosil GA, Schneider Gasser EM, Rudolph U, Keist R, Fritschy JM & Vogt KE.
(2006). Specific subtypes of GABAA receptors mediate phasic and tonic forms
of inhibition in hippocampal pyramidal neurons. J Neurophysiol 96, 846-857.
Pritchett DB, Sontheimer H, Shivers BD, Ymer S, Kettenmann H, Schofield PR &
Seeburg PH. (1989). Importance of a novel GABAA receptor subunit for
benzodiazepine pharmacology. Nature 338, 582-585.
Prys-Roberts C & Sear J. (1980). Steroid anaesthesia. Br J Anaesth 52, 363-365.
Puia G, Mienville JM, Matsumoto K, Takahata H, Watanabe H, Costa E & Guidotti A.
(2003). On the putative physiological role of allopregnanolone on GABAA
receptor function. Neuropharmacology 44, 49-55.
Puia G, Santi MR, Vicini S, Pritchett DB, Purdy RH, Paul SM, Seeburg PH & Costa E.
(1990). Neurosteroids act on recombinant human GABAA receptors. Neuron 4,
759-765.
Purdy RH, Moore PH, Jr., Rao PN, Hagino N, Yamaguchi T, Schmidt P, Rubinow DR,
Morrow AL & Paul SM. (1990a). Radioimmunoassay of 3 -hydroxy-5 -
pregnan-20-one in rat and human plasma. Steroids 55, 290-296.
Purdy RH, Morrow AL, Blinn JR & Paul SM. (1990b). Synthesis, metabolism, and
pharmacological activity of 3 -hydroxy steroids which potentiate GABA-
receptor-mediated chloride ion uptake in rat cerebral cortical
synaptoneurosomes. J Med Chem 33, 1572-1581.
Purdy RH, Morrow AL, Moore PH, Jr. & Paul SM. (1991). Stress-induced elevations of
-aminobutyric acid type A receptor-active steroids in the rat brain. Proc Natl
Acad Sci U S A 88, 4553-4557.
Ramos B, Lopez-Tellez JF, Vela J, Baglietto-Vargas D, del Rio JC, Ruano D, Gutierrez
A & Vitorica J. (2004). Expression of 5 GABAA receptor subunit in developing
rat hippocampus. Brain Res Dev Brain Res 151, 87-98.
References 198
Reddy DS & Rogawski MA. (2000). Chronic treatment with the neuroactive steroid
ganaxolone in the rat induces anticonvulsant tolerance to diazepam but not to
itself. J Pharmacol Exp Ther 295, 1241-1248.
Reddy DS & Rogawski MA. (2009). Neurosteroid replacement therapy for catamenial
epilepsy. Neurotherapeutics 6, 392-401.
Reddy DS & Rogawski MA. (2010). Ganaxolone suppression of behavioral and
electrographic seizures in the mouse amygdala kindling model. Epilepsy Res
89, 254-260.
Reith CA & Sillar KT. (1997). Pre- and postsynaptic modulation of spinal GABAergic
neurotransmission by the neurosteroid, 5 -pregnan-3 -ol-20-one. Brain Res
770, 202-212.
Reynolds DS, McKernan RM & Dawson GR. (2001). Anxiolytic-like action of diazepam:
which GABAA receptor subtype is involved? Trends Pharmacol Sci 22, 402-403.
Reynolds LM, Engin E, Tantillo G, Lau HM, Muschamp JW, Carlezon WA, Jr. &
Rudolph U. (2012). Differential roles of GABAA receptor subtypes in
benzodiazepine-induced enhancement of brain-stimulation reward.
Neuropsychopharmacology 37, 2531-2540.
Richerson GB. (2004). Looking for GABA in all the wrong places: the relevance of
extrasynaptic GABAA receptors to epilepsy. Epilepsy Curr 4, 239-242.
Rodgers RJ & Johnson NJ. (1995). Factor analysis of spatiotemporal and ethological
measures in the murine elevated plus-maze test of anxiety. Pharmacol Biochem
Behav 52, 297-303.
Rodgers RJ & Johnson NJ. (1998). Behaviorally selective effects of neuroactive
steroids on plus-maze anxiety in mice. Pharmacol Biochem Behav 59, 221-232.
Romeo E, Strohle A, Spalletta G, di Michele F, Hermann B, Holsboer F, Pasini A &
Rupprecht R. (1998). Effects of antidepressant treatment on neuroactive
steroids in major depression. Am J Psychiatry 155, 910-913.
Rudolph U, Crestani F, Benke D, Brunig I, Benson JA, Fritschy JM, Martin JR,
Bluethmann H & Mohler H. (1999). Benzodiazepine actions mediated by
specific -aminobutyric acidA receptor subtypes. Nature 401, 796-800.
Ruiz A, Walker MC, Fabian-Fine R & Kullmann DM. (2004). Endogenous Zinc inhibits
GABAA receptors in a hippocampal pathway. J Neurophysiol 91, 1091-1096.
Rupprecht R & Holsboer F. (1999). Neuroactive steroids: mechanisms of action and
neuropsychopharmacological perspectives. Trends Neurosci 22, 410-416.
References 199
Saalmann YB, Kirkcaldie MT, Waldron S & Calford MB. (2007). Cellular distribution of
the GABAA receptor-modulating 3 -hydroxy, 5 -reduced pregnane steroids in
the adult rat brain. J Neuroendocrinol 19, 272-284.
Sanacora G, Mason GF, Rothman DL, Behar KL, Hyder F, Petroff OA, Berman RM,
Charney DS & Krystal JH. (1999). Reduced cortical -aminobutyric acid levels in
depressed patients determined by proton magnetic resonance spectroscopy.
Arch Gen Psychiatry 56, 1043-1047.
Santi MR, Vicini S, Eldadah B & Neale JH. (1994). Analysis by polymerase chain
reaction of 1 and 6 GABAA receptor subunit mRNAs in individual cerebellar
neurons after whole-cell recordings. J Neurochem 63, 2357-2360.
Sarviharju M, Hyytia P, Hervonen A, Jaatinen P, Kiianmaa K & Korpi ER. (2006).
Lifelong ethanol consumption and brain regional GABAA receptor subunit
mRNA expression in alcohol-preferring rats. Alcohol 40, 159-166.
Scimemi A, Semyanov A, Sperk G, Kullmann DM & Walker MC. (2005). Multiple and
plastic receptors mediate tonic GABAA receptor currents in the hippocampus. J
Neurosci 25, 10016-10024.
Semyanov A, Walker MC & Kullmann DM. (2003). GABA uptake regulates cortical
excitability via cell type-specific tonic inhibition. Nat Neurosci 6, 484-490.
Semyanov A, Walker MC, Kullmann DM & Silver RA. (2004). Tonically active GABAA
receptors: modulating gain and maintaining the tone. Trends Neurosci 27, 262-
269.
Serra M, Pisul MG, Dazzi L, Purdy RH & Biggio G. (2002). Prevention of the stress-
induced increase in the concentration of neuroactive steroids in rat brain by
long-term administration of mirtazapine but not of fluoxetine. J
Psychopharmacol 16, 133-138.
Sheline YI. (2003). Neuroimaging studies of mood disorder effects on the brain. Biol
Psychiatry 54, 338-352.
Shen H, Gong QH, Aoki C, Yuan M, Ruderman Y, Dattilo M, Williams K & Smith SS.
(2007). Reversal of neurosteroid effects at 4 2 GABAA receptors triggers
anxiety at puberty. Nat Neurosci 10, 469-477.
Shen Q, Lal R, Luellen BA, Earnheart JC, Andrews AM & Luscher B. (2010). -
aminobutyric acid-type A receptor deficits cause hypothalamic-pituitary-adrenal
axis hyperactivity and antidepressant drug sensitivity reminiscent of melancholic
forms of depression. Biol Psychiatry 68, 512-520.
References 200
Shirayama Y & Chaki S. (2006). Neurochemistry of the nucleus accumbens and its
relevance to depression and antidepressant action in rodents. Curr
Neuropharmacol 4, 277-291.
Shirayama Y, Muneoka K, Fukumoto M, Tadokoro S, Fukami G, Hashimoto K & Iyo M.
(2011). Infusions of allopregnanolone into the hippocampus and amygdala, but
not into the nucleus accumbens and medial prefrontal cortex, produce
antidepressant effects on the learned helplessness rats. Hippocampus 21,
1105-1113.
Sieghart W & Sperk G. (2002). Subunit composition, distribution and function of GABAA
receptor subtypes. Curr Top Med Chem 2, 795-816.
Smart TG. (1997). Regulation of excitatory and inhibitory neurotransmitter-gated ion
channels by protein phosphorylation. Curr Opin Neurobiol 7, 358-367.
Smart TG, Hosie AM & Miller PS. (2004). Zn2+ ions: modulators of excitatory and
inhibitory synaptic activity. Neuroscientist 10, 432-442.
Smith KS & Rudolph U. (2012). Anxiety and depression: mouse genetics and
pharmacological approaches to the role of GABAA receptor subtypes.
Neuropharmacology 62, 54-62.
Smith SS, Gong QH, Hsu FC, Markowitz RS, ffrench-Mullen JM & Li X. (1998). GABAA
receptor 4 subunit suppression prevents withdrawal properties of an
endogenous steroid. Nature 392, 926-930.
Somogyi P, Takagi H, Richards JG & Mohler H. (1989). Subcellular localization of
benzodiazepine/GABAA receptors in the cerebellum of rat, cat, and monkey
using monoclonal antibodies. J Neurosci 9, 2197-2209.
Song M & Messing RO. (2005). Protein kinase C regulation of GABAA receptors. Cell
Mol Life Sci 62, 119-127.
Sperk G, Schwarzer C, Tsunashima K, Fuchs K & Sieghart W. (1997). GABAA receptor
subunits in the rat hippocampus I: Immunocytochemical distribution of 13
subunits. Neuroscience 80, 987-1000.
Stanford SC. (2007). The open field test: reinventing the wheel. J Psychopharmacol 21,
134-135.
Steland A, Padmanabhan AR & Akram M. (2011). Resampling methods for the
nonparametric and generalised Behrens-Fisher problems. Indian J Statistics 73-
A, 267-302.
Stell BM, Brickley SG, Tang CY, Farrant M & Mody I. (2003). Neuroactive steroids
reduce neuronal excitability by selectively enhancing tonic inhibition mediated
References 201
by subunit-containing GABAA receptors. Proc Natl Acad Sci U S A 100,
14439-14444.
Steru L, Chermat R, Thierry B & Simon P. (1985). The tail suspension test: a new
method for screening antidepressants in mice. Psychopharmacology (Berl) 85,
367-370.
Stocco DM & Clark BJ. (1996). Role of the steroidogenic acute regulatory protein
(StAR) in steroidogenesis. Biochem Pharmacol 51, 197-205.
Stoffel-Wagner B. (2003). Neurosteroid biosynthesis in the human brain and its clinical
implications. Ann N Y Acad Sci 1007, 64-78.
Strohle A, Pasini A, Romeo E, Hermann B, Spalletta G, di Michele F, Holsboer F &
Rupprecht R. (2000). Fluoxetine decreases concentrations of 3 ,5 -tetra-
hydro-deoxy-corticosterone (THDOC) in major depression. J Psychiatr Res 34,
183-186.
Strohle A, Romeo E, Hermann B, Pasini A, Spalletta G, di Michele F, Holsboer F &
Rupprecht R. (1999). Concentrations of 3 -reduced neuroactive steroids and
their precursors in plasma of patients with major depression and after clinical
recovery. Biol Psychiatry 45, 274-277.
Strous RD, Maayan R & Weizman A. (2006). The relevance of neurosteroids to clinical
psychiatry: from the laboratory to the bedside. Eur Neuropsychopharmacol 16,
155-169.
Szabadics J, Varga C, Molnar G, Olah S, Barzo P & Tamas G. (2006). Excitatory effect
of GABAergic axo-axonic cells in cortical microcircuits. Science 311, 233-235.
Tan KR, Brown M, Labouebe G, Yvon C, Creton C, Fritschy JM, Rudolph U & Luscher
C. (2010). Neural bases for addictive properties of benzodiazepines. Nature
463, 769-774.
Tan KR, Rudolph U & Luscher C. (2011). Hooked on benzodiazepines: GABAA
receptor subtypes and addiction. Trends Neurosci 34, 188-197.
Thierry B, Steru L, Simon P & Porsolt RD. (1986). The tail suspension test: ethical
considerations. Psychopharmacology (Berl) 90, 284-285.
Thomas P, Mortensen M, Hosie AM & Smart TG. (2005). Dynamic mobility of functional
GABAA receptors at inhibitory synapses. Nat Neurosci 8, 889-897.
Thomas P & Smart TG. (2005). HEK293 cell line: a vehicle for the expression of
recombinant proteins. J Pharmacol Toxicol Methods 51, 187-200.
References 202
Thompson SA, Whiting PJ & Wafford KA. (1996). Barbiturate interactions at the human
GABAA receptor: dependence on receptor subunit combination. Br J Pharmacol
117, 521-527.
Tobler I, Kopp C, Deboer T & Rudolph U. (2001). Diazepam-induced changes in sleep:
role of the 1 GABAA receptor subtype. Proc Natl Acad Sci U S A 98, 6464-
6469.
Trapani G, Franco M, Ricciardi L, Latrofa A, Genchi G, Sanna E, Tuveri F, Cagetti E,
Biggio G & Liso G. (1997). Synthesis and binding affinity of 2-
phenylimidazo[1,2- ]pyridine derivatives for both central and peripheral
benzodiazepine receptors. A new series of high-affinity and selective ligands for
the peripheral type. J Med Chem 40, 3109-3118.
Tretter V, Ehya N, Fuchs K & Sieghart W. (1997). Stoichiometry and assembly of a
recombinant GABAA receptor subtype. J Neurosci 17, 2728-2737.
Tretter V, Jacob TC, Mukherjee J, Fritschy JM, Pangalos MN & Moss SJ. (2008). The
clustering of GABAA receptor subtypes at inhibitory synapses is facilitated via
the direct binding of receptor 2 subunits to gephyrin. J Neurosci 28, 1356-
1365.
Tretter V, Kerschner B, Milenkovic I, Ramsden SL, Ramerstorfer J, Saiepour L, Maric
HM, Moss SJ, Schindelin H, Harvey RJ, Sieghart W & Harvey K. (2011).
Molecular basis of the -aminobutyric acid A receptor 3 subunit interaction with
the clustering protein gephyrin. J Biol Chem 286, 37702-37711.
Turkmen S, Backstrom T, Wahlstrom G, Andreen L & Johansson IM. (2011). Tolerance
to allopregnanolone with focus on the GABA-A receptor. Br J Pharmacol 162,
311-327.
Twelvetrees AE, Yuen EY, Arancibia-Carcamo IL, MacAskill AF, Rostaing P, Lumb MJ,
Humbert S, Triller A, Saudou F, Yan Z & Kittler JT. (2010). Delivery of
GABAARs to synapses is mediated by HAP1-KIF5 and disrupted by mutant
huntingtin. Neuron 65, 53-65.
Uzunov DP, Cooper TB, Costa E & Guidotti A. (1996). Fluoxetine-elicited changes in
brain neurosteroid content measured by negative ion mass fragmentography.
Proc Natl Acad Sci U S A 93, 12599-12604.
Uzunova V, Sheline Y, Davis JM, Rasmusson A, Uzunov DP, Costa E & Guidotti A.
(1998). Increase in the cerebrospinal fluid content of neurosteroids in patients
with unipolar major depression who are receiving fluoxetine or fluvoxamine.
Proc Natl Acad Sci U S A 95, 3239-3244.
References 203
Vallee M, Rivera JD, Koob GF, Purdy RH & Fitzgerald RL. (2000). Quantification of
neurosteroids in rat plasma and brain following swim stress and
allopregnanolone administration using negative chemical ionization gas
chromatography/mass spectrometry. Anal Biochem 287, 153-166.
van der Heyden JA, Molewijk E & Olivier B. (1987). Strain differences in response to
drugs in the tail suspension test for antidepressant activity.
Psychopharmacology (Berl) 92, 127-130.
van Rijnsoever C, Tauber M, Choulli MK, Keist R, Rudolph U, Mohler H, Fritschy JM &
Crestani F. (2004). Requirement of 5-GABAA receptors for the development of
tolerance to the sedative action of diazepam in mice. J Neurosci 24, 6785-6790.
VanDoren MJ, Matthews DB, Janis GC, Grobin AC, Devaud LL & Morrow AL. (2000).
Neuroactive steroid 3 -hydroxy-5 -pregnan-20-one modulates
electrophysiological and behavioral actions of ethanol. J Neurosci 20, 1982-
1989.
Vollenweider I, Smith KS, Keist R & Rudolph U. (2011). Antidepressant-like properties
of 2-containing GABAA receptors. Behav Brain Res 217, 77-80.
Wang H, Bedford FK, Brandon NJ, Moss SJ & Olsen RW. (1999). GABAA-receptor-
associated protein links GABAA receptors and the cytoskeleton. Nature 397, 69-
72.
Wang M, He Y, Eisenman LN, Fields C, Zeng CM, Mathews J, Benz A, Fu T, Zorumski
E, Steinbach JH, Covey DF, Zorumski CF & Mennerick S. (2002). 3 -
hydroxypregnane steroids are pregnenolone sulfate-like GABAA receptor
antagonists. J Neurosci 22, 3366-3375.
Whiting P, McKernan RM & Iversen LL. (1990). Another mechanism for creating
diversity in -aminobutyrate type A receptors: RNA splicing directs expression of
two forms of 2 subunit, one of which contains a protein kinase C
phosphorylation site. Proc Natl Acad Sci U S A 87, 9966-9970.
Wieland HA, Luddens H & Seeburg PH. (1992). A single histidine in GABAA receptors
is essential for benzodiazepine agonist binding. J Biol Chem 267, 1426-1429.
Wieland S, Lan NC, Mirasedeghi S & Gee KW. (1991). Anxiolytic activity of the
progesterone metabolite 5 -pregnan-3 -o1-20-one. Brain Res 565, 263-268.
Wilkins ME, Hosie AM & Smart TG. (2002). Identification of a subunit TM2 residue
mediating proton modulation of GABA type A receptors. J Neurosci 22, 5328-
5333.
References 204
Wilkins ME, Hosie AM & Smart TG. (2005). Proton modulation of recombinant GABAA
receptors: influence of GABA concentration and the subunit TM2-TM3
domain. J Physiol 567, 365-377.
Winter L, Nadeson R, Tucker AP & Goodchild CS. (2003). Antinociceptive properties of
neurosteroids: a comparison of alphadolone and alphaxalone in potentiation of
opioid antinociception. Anesth Analg 97, 798-805.
Wisden W, Cope D, Klausberger T, Hauer B, Sinkkonen ST, Tretter V, Lujan R, Jones
A, Korpi ER, Mody I, Sieghart W & Somogyi P. (2002). Ectopic expression of
the GABAA receptor 6 subunit in hippocampal pyramidal neurons produces
extrasynaptic receptors and an increased tonic inhibition. Neuropharmacology
43, 530-549.
Wisden W, Laurie DJ, Monyer H & Seeburg PH. (1992). The distribution of 13 GABAA
receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon,
mesencephalon. J Neurosci 12, 1040-1062.
Wisden W & Seeburg PH. (1992). GABAA receptor channels: from subunits to
functional entities. Curr Opin Neurobiol 2, 263-269.
Wu X, Wu Z, Ning G, Guo Y, Ali R, Macdonald RL, De Blas AL, Luscher B & Chen G.
(2012). -aminobutyric acid type A (GABAA) receptor subunits play a direct
role in synaptic versus extrasynaptic targeting. J Biol Chem 287, 27417-27430.
Xi ZX, Ramamoorthy S, Shen H, Lake R, Samuvel DJ & Kalivas PW. (2003). GABA
transmission in the nucleus accumbens is altered after withdrawal from
repeated cocaine. J Neurosci 23, 3498-3505.
Yamada K, Watanabe A, Iwayama-Shigeno Y & Yoshikawa T. (2003). Evidence of
association between gamma-aminobutyric acid type A receptor genes located
on 5q34 and female patients with mood disorders. Neurosci Lett 349, 9-12.
Yee BK, Keist R, von Boehmer L, Studer R, Benke D, Hagenbuch N, Dong Y, Malenka
RC, Fritschy JM, Bluethmann H, Feldon J, Mohler H & Rudolph U. (2005). A
schizophrenia-related sensorimotor deficit links 3-containing GABAA receptors
to a dopamine hyperfunction. Proc Natl Acad Sci U S A 102, 17154-17159.
Zeilhofer HU. (2008). Loss of glycinergic and GABAergic inhibition in chronic pain -
contributions of inflammation and microglia. Int Immunopharmacol 8, 182-187.
Zeilhofer HU, Wildner H & Yevenes GE. (2012). Fast synaptic inhibition in spinal
sensory processing and pain control. Physiol Rev 92, 193-235.
Zhu WJ & Vicini S. (1997). Neurosteroid prolongs GABAA channel deactivation by
altering kinetics of desensitized states. J Neurosci 17, 4022-4031.